• Vita
  • Publikationen
  • Projekte

Stefan Wesarg studierte Physik an der Technischen Universität Berlin, der Ecole Nationale Supérieure de Physique de Marseille in Frankreich sowie der Universität Heidelberg. Themenschwerpunkte seines Studiums waren Astrophysik sowie Medizinische Physik. Seine Diplomarbeit machte er 2001 am Deutschen Krebsforschungszentrum (DKFZ) in Heidelberg.

Von 2001 bis 2007 war er wissenschaftlicher Mitarbeiter am Fraunhofer IGD in Darmstadt in der Abteilung Cognitive Computing & Medical Imaging (stellv. Abteilungsleiter von 2006 bis 2007). Dort arbeitete er in verschiedenen Projekten im Bereich der medizinischen Bildverarbeitung: Ultraschall, medizinische Navigationssysteme sowie Analyse von Herzbilddaten.

Im Jahr 2007 wurde er vom Fachbereich Informatik der Technischen Universität Darmstadt promoviert (Dr.-Ing.). Anschließend arbeitete er von 2007 bis Ende 2011 als Postdoc an der Technischen Universität Darmstadt und baute die neue Nachwuchsforschergruppe Medical Computing auf. Als Leiter dieser Gruppe warb er erfolgreich Forschungsförderung bei der Deutschen Forschungsgemeinschaft (DFG) ein.

Seit 2012 ist er zurück am Fraunhofer IGD und Leiter der Abteilung Visual Healthcare Technologies. Fokus der aktuellen Arbeiten sind bildgestützte Eingriffe und Chirurgie, Ultraschall-Anwendungen für Diagnose und Planung, statistische Modellierung von Organen sowie Simulation und Planung für medizinischer Anwendungen.

Stefan Wesarg wurde mit dem »Eurographics 2005 Medical Prize, 2nd« für seinen Beitrag »MEDARPA – An Augmented Reality System for Supporting Minimally Invasive Interventions« ausgezeichnet. Sein Paper »Localizing Calcifications in Cardiac CT Data Sets Using a New Vessel Segmentation Approach« mit dem »Best Paper 2006, 2nd« des Journal of Digital Imaging bedacht. Er arbeitet regelmäßig als Reviewer verschiedener peer-reviewed Journals (z.B. IEEE Trans Med Img, Int. Journal for CARS, Journal of Digital Imaging) als auch für verschiedene Konferenzen und Workshops (z.B. IEEE ISBI, MICCAI Workshops).

Show publication details

Oyarzun Laura, Cristina; Drechsler, Klaus; Wesarg, Stefan; Bale, Reto

Accurate Physics-Based Registration for the Outcome Validation of Minimal Invasive Interventions and Open Liver Surgeries

2017

IEEE Transactions on Biomedical Engineering, Vol.64 (2017), 2, pp. 362-371

The purpose of this paper is to present an outcome validation tool for tumor radiofrequency (RF) ablation and resection. Methods: Intervention assessment tools require an accurate registration of both pre- and postoperative computed tomographies able to handle big deformations. Therefore, a physics-based method is proposed with that purpose. To increase the accuracy both automatically detected internal and surface physical landmarks are incorporated in the registration process. Results: The algorithm has been evaluated in 25 clinical datasets containing RF ablations, resections, and patients with recurrent tumors. The achieved accuracy is 1.2 mm measured as mean internal distance between vessel landmarks and a positive predictive value of 0.95. The quantitative and qualitative results of the outcome validation tool show that in 50% of the cases tumors were only partially covered by the treatment. Conclusion: The use of internal and surface landmarks combined with a physics-based registration method increases the accuracy of the results compared to the accuracy of state of the art methods. An accurate outcome validation tool is important in order to certify that the tumor and its safety margin were fully covered by the treatment. Significance: An accurate outcome validation tool can result in a decrease of the tumor recurrence rate.

Show publication details

Jung, Florian; Biebl-Rydlo, Medea; Daisne, Jean-François; Wesarg, Stefan

Automatic Sentinel Lymph Node Localization in Head and Neck Cancer Using a Coupled Shape Model Algorithm

2017

Cardoso, Jorge M. (Ed.) et al.: Computer Assisted and Robotic Endoscopy and Clinical Image-Based Procedures : 4th International Workshop, CARE 2017 and 6th International Workshop, CLIP 2017. Held in Conjunction with MICCAI 2017.. Berlin, Heidelberg, New York: Springer, 2017. (Lecture Notes in Computer Science (LNCS) 10550), pp. 133-140

The localization and analysis of the sentinel lymph node for patients diagnosed with cancer, has significant influence on the prognosis, outcome and treatment of the disease. We present a fully automatic approach to localize the sentinel lymph node and additional active nodes and determine their lymph node level on SPECT-CT data. This is a crucial prerequisite for the planning of radiation therapy or a surgical neck dissection. Our approach was evaluated on 17 lymph nodes. The detection rate of the lymph nodes was 94%; and 88% of the lymph nodes were correctly assigned to their corresponding lymph node level. The proposed algorithm targets a very important topic in clinical practice. The first results are already very promising. The next step has to be the evaluation on a larger data set.

Show publication details

Raudaschl, Patrik F.; Zaffino, Paolo; Spadea, Maria Francesca; Chen, Antong; Dawant, Benoit M.; Albrecht, Thomas; Gass, Tobias; Langgut, Christoph; Lüthi, Marcel; Jung, Florian; Knapp, Oliver; Wesarg, Stefan; Mannion-Haworth, Richard; Bowes, Mike; Ashman, Annaliese; Guillard, Gwenael; Brett, Alan; Vincent, Graham; Orbes-Arteaga, Mauricio; Cárdenas-Peña, David; Castellanos-Dominguez, German; Aghdasi, Nava; Li, Yangming; Berens, Angelique; Moe, Kris; Hannaford, Blake; Schubert, Rainer; Fritscher, Karl D.

Evaluation of Segmentation Methods on Head and Neck CT: Auto-segmentation Challenge 2015

2017

Medical Physics, Vol.44 (2017), 5, pp. 2020-2036

Purpose Automated delineation of structures and organs is a key step in medical imaging. However, due to the large number and diversity of structures and the large variety of segmentation algorithms, a consensus is lacking as to which automated segmentation method works best for certain applications. Segmentation challenges are a good approach for unbiased evaluation and comparison of segmentation algorithms. Methods In this work, we describe and present the results of the Head and Neck Auto-Segmentation Challenge 2015, a satellite event at the Medical Image Computing and Computer Assisted Interventions (MICCAI) 2015 conference. Six teams participated in a challenge to segment nine structures in the head and neck region of CT images: brainstem, mandible, chiasm, bilateral optic nerves, bilateral parotid glands, and bilateral submandibular glands. Results This paper presents the quantitative results of this challenge using multiple established error metrics and a well-defined ranking system. The strengths and weaknesses of the different auto-segmentation approaches are analyzed and discussed. Conclusions The Head and Neck Auto-Segmentation Challenge 2015 was a good opportunity to assess the current state-of-the-art in segmentation of organs at risk for radiotherapy treatment. Participating teams had the possibility to compare their approaches to other methods under unbiased and standardized circumstances. The results demonstrate a clear tendency toward more general purpose and fewer structure-specific segmentation algorithms.

Show publication details

Ma, Jingting; Wang, Anqi; Lin, Feng; Wesarg, Stefan; Erdt, Marius

Nonlinear Statistical Shape Modeling for Ankle Bone Segmentation Using a Novel Kernelized Robust PCA

2017

Descoteaux, Maxime (Ed.) et al.: Medical Image Computing and Computer Assisted Intervention - MICCAI 2017: Part I : 20th International Conference. Springer International Publishing, 2017. (Lecture Notes in Computer Science (LNCS) 10433), pp.136-143

International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI) <20, 2017, Quebec City, Canada>

Statistical shape models (SSMs) are widely employed in medical image segmentation. However, an inferior SSM will degenerate the quality of segmentations. It is challenging to derive an efficient model because: (1) often the training datasets are corrupted by noise and/or artifacts; (2) conventional SSM is not capable to capture nonlinear variabilities of a population of shape. Addressing these challenges, this work aims to create SSMs that are not only robust to abnormal training data but also satisfied with nonlinear distribution. As Robust PCA is an efficient tool to seek a clean low-rank linear subspace, a novel kernelized Robust PCA (KRPCA) is proposed to cope with nonlinear distribution for statistical shape modeling. In evaluation, the built nonlinear model is used in ankle bone segmentation where 9 bones are separately distributed. Evaluation results show that the model built with KRPCA has a significantly higher quality than other state-of-the-art methods.

Show publication details

Prinold, Joe A. I.; Mazzà, Claudia; Di Marco, Roberto; Hannah, Iain; Malattia, Clara; Magni-Manzoni, Silvia; Petrarca, Maurizio; Ronchetti, Anna B.; Tanturri De Horatio, Laura; van Dijkhuizen, Pieter E.H.; Wesarg, Stefan; Viceconti, Marco

A Patient-Specific Foot Model for the Estimate of Ankle Joint Forces in Patients with Juvenile Idiopathic Arthritis

2016

Annals of Biomedical Engineering, Vol.44 (2016), 1, pp. 247-257. First online: 15 September 2015

Juvenile idiopathic arthritis (JIA) is the leading cause of childhood disability from a musculoskeletal disorder. It generally affects large joints such as the knee and the ankle, often causing structural damage. Different factors contribute to the damage onset, including altered joint loading and other mechanical factors, associated with pain and inflammation. The prediction of patients' joint loading can hence be a valuable tool in understanding the disease mechanisms involved in structural damage progression. A number of lower-limb musculoskeletal models have been proposed to analyse the hip and knee joints, but juvenile models of the foot are still lacking. This paper presents a modelling pipeline that allows the creation of juvenile patient-specific models starting from lower limb kinematics and foot and ankle MRI data. This pipeline has been applied to data from three children with JIA and the importance of patient-specific parameters and modelling assumptions has been tested in a sensitivity analysis focused on the variation of the joint reaction forces. This analysis highlighted the criticality of patient-specific definition of the ankle joint axes and location of the Achilles tendon insertions. Patient-specific detection of the Tibialis Anterior, Tibialis Posterior, and Peroneus Longus origins and insertions were also shown to be important.

Show publication details

Noll, Matthias; Wesarg, Stefan

An Automatic Free Fluid Detection for Morrison´s-Pouch

2016

Shekhar, Raj (Ed.) et al.: Clinical Image-Based Procedures. Translational Research in Medical Imaging : 5th International Workshop, CLIP 2016. Held in Conjunction with MICCAI 2016. Berlin, Heidelberg, New York: Springer, 2016. (Lecture Notes in Computer Science (LNCS) 9958), pp. 77-84

International Workshop on Clinical Image-based Procedures (CLIP) <5, 2016, Athens, Greece>

Ultrasound provides a useful and readily available imaging tool to detect free fluids in blunt abdominal trauma patients. However, applying conventional 2D ultrasound to diagnose the patient requires a well trained physician. In this paper we describe a fully automatic free fluid detection pipeline for the hepathorenal recess or Morrison's pouch using 3D ultrasound acquisitions. The image data is collected using the standardized "Focused Assessment with Sonography for Trauma" (FAST) exam. Our method extracts key structures like the kidney and the liver from the image data and uses their relative positions to search and detect free fluids between the organ interfaces. To evaluate our method we have developed a free fluid simulation that allows us to generate free fluid images using acquisitions of healthy volunteers. Our intentions are to enable even untrained ultrasound operators to perform a free fluid diagnosis of an injured person. In order to do this, our method additionally provides basic image acquisition guidance information.

Show publication details

Noll, Matthias; Nadolny, Anne; Wesarg, Stefan

Automated Kidney Detection for 3D Ultrasound Using Scan Line Searching

2016

The International Society for Optical Engineering (SPIE): Proceedings of SPIE, Article 9790-46, 6 p.

SPIE Medical Imaging Symposium <2016, San Diego, CA, USA>

Ultrasound (U/S) is a fast and non-expensive imaging modality that is used for the examination of various anatomical structures, e.g. the kidneys. One important task for automatic organ tracking or computer-aided diagnosis is the identification of the organ region. During this process the exact information about the transducer location and orientation is usually unavailable. This renders the implementation of such automatic methods exceedingly challenging. In this work we like to introduce a new automatic method for the detection of the kidney in 3D U/S images. This novel technique analyses the U/S image data along virtual scan lines. Here, characteristic texture changes when entering and leaving the symmetric tissue regions of the renal cortex are searched for. A subsequent feature accumulation along a second scan direction produces a 2D heat map of renal cortex candidates, from which the kidney location is extracted in two steps. First, the strongest candidate as well as its counterpart are extracted by heat map intensity ranking and renal cortex size analysis. This process exploits the heat map gap caused by the renal pelvis region. Substituting the renal pelvis detection with this combined cortex tissue feature increases the detection robustness. In contrast to model based methods that generate characteristic pattern matches, our method is simpler and therefore faster. An evaluation performed on 61 3D U/S data sets showed, that in 55 cases showing none or minor shadowing the kidney location could be correctly identified.

Show publication details

Jung, Florian; Knapp, Oliver; Wesarg, Stefan

Automatic Segmentation of Structures in CT Head and Neck Images using a Coupled Shape Model

2016

MIDAS Journal, (2016), 3 p.

The common approach to do a fully automatic segmentation of multiple struc tures is an atlas or multi-atlas based solution. These already have proven to be suitable for the segmentation of structures in the head and neck area and provide very accurate segmentation results, but can struggle with challenging cases with unnatural postures, where the registration of the reference patient(s) is extremely difficult. Therefore, we propose an coupled shape model (CoSMo) algorithm for the segmentation relevant structures in parallel. The model adaptation to a test image is done with respect to the appearance of its items and the trained articulation space. Even on very challenging data sets with unnatural postures, which occur far more often than expected, the model adaptation algorithm succeeds. The approach is based on an articulated atlas cite{Steger2012a}, that is trained from a set of manually labeled training samples. Furthermore, we have combined the initial solution with statistical shape models cite{Kirschner2011} to represent structures with high shape variation. CoSMo is not tailored to specifc structures or regions. It can be trained from any set of given gold standard segmentations and makes it thereby very generic.

Show publication details

Shekhar, Raj; Wesarg, Stefan; González Ballester, Miguel Angel; Drechsler, Klaus; Sato, Yoshinobu; Erdt, Marius; Linguraru, Marius George; Oyarzun Laura, Cristina

Clinical Image-Based Procedures. Translational Research in Medical Imaging: 5th International Workshop, CLIP 2016. Held in Conjunction with MICCAI 2016

2016

Berlin, Heidelberg, New York : Springer, 2016

International Workshop on Clinical Image-based Procedures (CLIP) <5, 2016, Athens, Greece>

Lecture Notes in Computer Science (LNCS) 9958

Show publication details

Oyarzun Laura, Cristina; Oelmann, Simon; Drechsler, Klaus; Wesarg, Stefan

Reducing Over- and Undersegmentations of the Liver in Computed Tomographies Using Anatomical Knowledge

2016

Efthyvoulos, Kyriacou (Ed.) et al.: XIV Mediterranean Conference on Medical and Biological Engineering and Computing : MEDICON 2016. Berlin, Heidelberg, New York: Springer, 2016. (IFMBE Proceedings 57), pp. 382-387

Mediterranean Conference on Medical and Biological Engineering and Computing (MEDICON) <14, 2016, Paphos, Cyprus>

In the last decades several liver segmentation methods have been proposed. The proposed methods go from region growing to the more complex statistical shape models. Despite the robustness of those algorithms, liver segmentation is still a challenging task especially in areas in which its neighboring organs have similar intensities, e.g., heart and ribcage. In addition to this, pathological organs that contain tumors near their surface present additional difficulties. This paper presents a solution to increase the accuracy of those algorithms in the aforementioned areas. The effect of the improvement using the generated heart and ribcage walls (7% and 1% respectively) is evaluated on 9 clinical computer tomographies (CT). The improvement (12 %) when tumors are near the surface, on the contrary, is tested on 7 clinical CT images.

Show publication details

Ma, Jingting; Lin, Feng; Honsdorf, Jonas; Lentzen, Katharina; Wesarg, Stefan; Erdt, Marius

Weighted Robust PCA for Statistical Shape Modeling

2016

Zheng, Guoyan (Ed.) et al.: Medical Imaging and Augmented Reality : 7th International Conference, MIAR 2016. Proceedings. Springer, 2016. (Lecture Notes in Computer Science (LNCS) 9805), pp. 343-353

International Conference on Medical Imaging and Augmented Reality (MIAR) <7, 2016, Bern, Switzerland>

Statistical shape models (SSMs) play an important role in medical image analysis. A sufficiently large number of high quality datasets is needed in order to create a SSM containing all possible shape variations. However, the available datasets may contain corrupted or missing data due to the fact that clinical images are often captured incompletely or contain artifacts. In this work, we propose a weighted Robust Principal Component Analysis (WRPCA) method to create SSMs from incomplete or corrupted datasets. In particular, we introduce a weighting scheme into the conventional Robust Principal Component Analysis (RPCA) algorithm in order to discriminate unusable data from meaningful ones in the decomposition of the training data matrix more accurately. For evaluation, the proposed WRPCA is compared with conventional RPCA on both corrupted (63 CT datasets of the liver) and incomplete datasets (15 MRI datasets of the human foot). The results show a significant improvement in terms of reconstruction accuracy on both datasets.

Show publication details

Oelmann, Simon; Oyarzun Laura, Cristina; Drechsler, Klaus; Wesarg, Stefan

Active Contour based Segmentation of Resected Livers in CT Images

2015

Ourselin, Sébastien (Ed.) et al.: Medical Imaging 2015: Image Processing. Bellingham: SPIE Press, 2015. (Proceedings of SPIE 9413), pp. 941316-1 - 941316-6

SPIE Medical Imaging Symposium <2015, Orlando, FL, USA>

The majority of state of the art segmentation algorithms are able to give proper results in healthy organs but not in pathological ones. However, many clinical applications require an accurate segmentation of pathological organs. The determination of the target boundaries for radiotherapy or liver volumetry calculations are examples of this. Volumetry measurements are of special interest after tumor resection for follow up of liver regrow. The segmentation of resected livers presents additional challenges that were not addressed by state of the art algorithms. This paper presents a snakes based algorithm specially developed for the segmentation of resected livers. The algorithm is enhanced with a novel dynamic smoothing technique that allows the active contour to propagate with different speeds depending on the intensities visible in its neighborhood. The algorithm is evaluated in 6 clinical CT images as well as 18 artificial datasets generated from additional clinical CT images.

Show publication details

Wichmann, Julian L.; Booz, Christian; Wesarg, Stefan; Bauer, Ralf W.; Kerl, J. Matthias; Fischer, Sebastian; Lehnert, Thomas; Vogl, Thomas J.; Khan, M. Fawad; Kafchitsas, Konstantinos

Quantitative Dual-energy CT for Phantomless Evaluation of Cancellous Bone Mineral Density of the Vertebral Pedicle: Correlation with Pedicle Screw Pull-out Strength

2015

European Radiology, Vol.25 (2015), 6, pp. 1714-1720. Published online 07. December 2014

Objectives: To evaluate quantitative dual-energy computed tomography (DECT) for phantomless analysis of cancellous bone mineral density (BMD) of vertebral pedicles and to assess the correlation with pedicle screw pull-out strength. Methods: Twenty-nine thoracic and lumbar vertebrae from cadaver specimens were examined with DECT. Using dedicated post-processing software, a pedicle screw vector was mapped (R1, intrapedicular segment of the pedicle vector; R2, intermediate segment; R3, intracorporal segment; global, all segments) and BMD was calculated. To invasively evaluate pedicle stability, pedicle screws were drilled through both pedicles and left pedicle screw pull-out strength was measured. Resulting values were correlated using the paired t test and Pearson's linear correlation. Results: Average pedicle screw vector BMD (R1, 0.232 g/cm3; R2, 0.166 g/cm3; R3, 0.173 g/cm3; global, 0.236 g/cm3) showed significant differences between R1-R2 (P<0.002) and R1-R3 (P<0.034) segments while comparison of R2-R3 did not reach significance (P>0.668). Average screw pull-out strength (639.2 N) showed a far stronger correlation with R1 (r=0.80; P<0.0001) than global BMD (r=0.42; P=0.025), R2 (r=0.37; P=0.048) and R3 (r=0.33; P=0.078) segments. Conclusions: Quantitative DECT allows for phantomless BMD assessment of the vertebral pedicle. BMD of the intrapedicular segment shows a significantly stronger correlation with pedicle screw pull-out strength than other segments.

Show publication details

Jung, Florian; Hilpert, Julia; Wesarg, Stefan

Segmentierung von zervikalen Lymphknoten in T1-gewichteten MRT-Aufnahmen

2015

Handels, H. (Ed.) et al.: Bildverarbeitung für die Medizin 2015 : Algorithmen - Systeme - Anwendungen. Proceedings des Workshops. Wiesbaden: Springer, Vieweg, 2015. (Informatik aktuell), pp. 353-358

Workshop Bildverarbeitung für die Medizin (BVM) <18, 2015, Lübeck, Germany>

Die Untersuchung von Größe und Aussehen eines Lymphknotens kann ein entscheidender Indikator für die Existenz eines Tumors sein und ist außerdem ein probates Mittel, um Verlaufsanalysen bei einem Patienten durchzuführen, welche wiederum maßgeblichen Einfluss auf die Behandlung haben können. Um die Größe und andere Parameter des Lymphknotens bestimmen zu können, ist zuerst eine Segmentierung vonnöten. Wir präsentieren ein neues Verfahren für die halbautomatische Segmentierung von Lymphknoten auf MR-Datensätzen. Unser Ansatz verwendet eine Wasserscheidentransformation als Grundlage und kombiniert diese mit einem Radialstrahlbasierten Verfahren, um eine möglichst akkurate Segmentierung des Lymphknotens zu erhalten. Für die Evaluation wurden 95 Lymphknoten-Segmentierungen aus 17 verschiedenen, kontrastverstärkten T1-gewichteten Patientendatensätzen verwendet. Das durchschnittliche Dice Ähnlichkeitsmaß lag bei 0.69}0.15 und die mittlere Oberflächendistanz bei 0.65}0.54mm.

Show publication details

Wang, Anqi; Noll, Matthias; Wesarg, Stefan

Tumorsegmentierung in CD3/CD8-gefärbten Histopathologien

2015

Handels, H. (Ed.) et al.: Bildverarbeitung für die Medizin 2015 : Algorithmen - Systeme - Anwendungen. Proceedings des Workshops. Wiesbaden: Springer, Vieweg, 2015. (Informatik aktuell), pp. 347-352

Workshop Bildverarbeitung für die Medizin (BVM) <18, 2015, Lübeck, Germany>

Segmentierung von bestimmten Gewebetypen in Histopathologien ist eine oft untersuchte Fragestellung. Üblicherweise werden dafür Gewebeproben mit Hämatoxylin-Eosin(HE)-Färbung verwendet. CD3/CD8-F¨arbungen hingegen sind nötig zur Sichtbarmachung von Immunzellen, differenzieren aber nur wenig zwischen unterschiedlichen Gewebearten. Vorteilhaft wäre es, wenn aus nur einem Gewebeschnitt mit einer bestimmten Färbung beide Informationen extrahiert werden könnten. In dieser Arbeit stellen wir ein Segmentierungsverfahren auf CD3/CD8-gef¨arbten Gewebeproben vor, das effizient zu berechnende und gleichzeitig aussagekräftige Features als Eingabe für einen Clustering- Algorithmus verwendet. In der Evaluation wird ein durchschnittlicher Accuracy-Wert von 94,44% erzielt. Dieser Wert ist vergleichbar mit den Ergebnissen verwandter State of the Art Methoden, die HE-gefärbte Proben einsetzen.

Show publication details

Wesarg, Stefan; Jung, Florian; Steger, Sebastian

Abbildung anatomischen Wissens: Ein Kopf-/Hals-Atlas multipler gekoppelter Strukturen

2014

Endoskopie heute, Vol.27 (2014), 3, pp. 141-145

Medizinische Bilddaten enthalten anatomische Informationen. Die Extraktion derselben durch manuelles Markieren ist unter Berücksichtigung der Datenmenge vor allem bei radiologischen 3D-Bilddaten nicht mehr vernünftig durchführbar. Hier helfen computerbasierte, automatische Verfahren. Nicht alle anatomischen Regionen heben sich durch deutliche Kontrastunterschiede von der Umgebung ab. Dennoch sind Radiologen in der Lage, auch solche Bereiche in den Bilddaten zuzuordnen. Für automatisch ablaufende Algorithmen besteht dabei die Herausforderung, das anatomische Wissen eines klinischen Experten in einer für den Computer verständlichen Form zu repräsentieren. Für den Kopf-/Hals-Bereich haben wir ein solches Modell entwickelt, das anatomische Strukturen enthält, die miteinander gekoppelt sind. Das Modell repräsentiert diese hinsichtlich ihrer Form und Größe als auch ihrer relativen Lage zueinander. Dabei wird zwischen aktiven und passiven Strukturen unterschieden. Erstere repräsentieren deutlich von der Umgebung abgrenzbare Bildbereiche, letztere Regionen, die nur aufgrund ihrer relativen Lage zu anderen Organen identifizierbar sind. Der neue Modellierungsansatz ist dabei einerseits flexibler als klassische atlasbasierte Segmentierungsverfahren. Andererseits sind das neue Verfahren und das dafür entwickelte Framework generisch und können mit geringen Adaptionen auch für Fragestellungen jenseits des Kopf-/Hals-Bereichs eingesetzt werden.

Show publication details

Oyarzun Laura, Cristina; Drechsler, Klaus; Wesarg, Stefan

Anatomical Discovery: Finding Organs in the Neighborhood of the Liver

2014

Romero, Laura M. Roa (Ed.): XIII Mediterranean Conference on Medical and Biological Engineering and Computing : MEDICON 2013. Berlin, Heidelberg, New York: Springer, 2014. (IFMBE Proceedings 41), pp. 348-351

The Mediterranean Conference on Medical and Biological Engineering and Computing (MEDICON) <13, 2013, Seville, Spain>

Image segmentation and registration algorithms are fundamental to assist medical doctors for better treatment of the patients. To this end accuracy in the results given by those algorithms is crucial. The surroundings of the organ to be segmented or registered can provide additional information that at the end improves the result. In this paper a novel algorithm to detect the organs that surround the liver is introduced. Even though our work is focused on the liver, the algorithm could be extended to other parts of the body. The algorithm has been tested in 24 clinical CT datasets. In addition to this, an example application is introduced for which the detection is a useful tool.

Show publication details

Noll, Matthias; Li, Xin; Wesarg, Stefan

Automated Kidney Detection and Segmentation in 3D Ultrasound

2014

Erdt, Marius (Ed.) et al.: Clinical Image-Based Procedures. Translational Research in Medical Imaging : Second International Workshop, CLIP 2013. Held in Conjunction with MICCAI 2013. Berlin, Heidelberg, New York: Springer, 2014. (Lecture Notes in Computer Science (LNCS) 8361), pp. 83-90

International Workshop on Clinical Image-based Procedures (CLIP) <2, 2013, Nagoya, Japan>

Ultrasound provides the physical capabilities for a fast and save disease diagnosis in various medical scenarios including renal exams and patient trauma assessment. However, the experience of the ultrasound operator is the key element in performing ultrasound diagnosis. Thus, we like to introduce our automatic kidney detection and segmentation algorithm for 3D ultrasound. The approach utilizes basic kidney shape information to detect the kidney position. Following, the Level Set algorithm is applied to segment the detection result. In combination this method may help physicians and inexperienced trainees to achieve kidney detection and segmentation for diagnostic purposes.

Show publication details

Erdt, Marius; Linguraru, Marius George; Oyarzun Laura, Cristina; Shekhar, Raj; Wesarg, Stefan; González Ballester, Miguel Angel; Drechsler, Klaus

Clinical Image-Based Procedures. Translational Research in Medical Imaging: Second International Workshop, CLIP 2013. Held in Conjunction with MICCAI 2013

2014

Berlin, Heidelberg, New York : Springer, 2014

International Workshop on Clinical Image-based Procedures (CLIP) <2, 2013, Nagoya, Japan>

Lecture Notes in Computer Science (LNCS) 8361

This book constitutes revised selected papers from the International Workshop on Clinical Image-Based Procedures, CLIP 2013, held in conjunction with MICCAI 2012 in Nagoya, Japan, in September 2013. The 19 papers presented in this volume were carefully reviewed and selected from 26 submissions. The workshop was a productive and exciting forum for the discussion and dissemination of clinically tested, state-of-the-art methods for image-based planning, monitoring and evaluation of medical procedures.

Show publication details

Linguraru, Marius George; Oyarzun Laura, Cristina; Shekhar, Raj; Wesarg, Stefan; González Ballester, Miguel Angel; Drechsler, Klaus; Sato, Yoshinobu; Erdt, Marius

Clinical Image-Based Procedures. Translational Research in Medical Imaging: Third International Workshop, CLIP 2014. Held in Conjunction with MICCAI 2014

2014

Berlin, Heidelberg, New York : Springer, 2014

International Workshop on Clinical Image-based Procedures (CLIP) <3, 2014, Boston, MA, USA>

Lecture Notes in Computer Science (LNCS) 8680

CLIP 2014 provided a successful forum for the dissemination of emerging image-based clinical techniques. Specific topics included pre-interventional image segmentation and classification (to support diagnosis and clinical decision making), interventional and surgical planning and analysis of dynamic images, and evaluation, visualization and correction techniques for image-based procedures. Clinical applications covered the skull and the brain, cardiac defects, blood vessels, abdominal organs, and cancer in adults and children. The presentations and discussions around the meeting emphasized current challenges and emerging techniques in image-based procedures, strategies for clinical translation of image-based techniques, the role of computational anatomy and image analysis for surgical planning and interventions, and the contribution of medical image analysis to open and minimally invasive surgery.

Show publication details

Jung, Florian; Steger, Sebastian; Knapp, Oliver; Noll, Matthias; Wesarg, Stefan

COSMO - Coupled Shape Model for Radiation Therapy Planning of Head and Neck Cancer

2014

Linguraru, Marius George (Ed.) et al.: Clinical Image-Based Procedures. Translational Research in Medical Imaging : Third International Workshop, CLIP 2014. Held in Conjunction with MICCAI 2014. Berlin, Heidelberg, New York: Springer, 2014. (Lecture Notes in Computer Science (LNCS) 8680), pp. 25-32

International Workshop on Clinical Image-based Procedures (CLIP) <3, 2014, Boston, MA, USA>

Radiation therapy plays a major role in head and neck cancer treatment. Segmentation of organs at risk prior to the radiation therapy helps to prevent the radiation beam from damaging healthy tissue, whereas a concentrated ray can target the cancerous regions. Unfortunately, the manual annotation of all relevant structures in the head and neck area is very time-consuming and existing atlas-based solutions don't provide sufficient segmentation accuracy. Therefore, we propose a coupled shape model (CoSMo) for the segmentation of key structures within the head and neck area. The model's adaptation to a test image is done with respect to the appearance of its items and the trained articulation space. 40 data sets labeled by clinicians containing 22 structures were used to build the CoSMo. Even on very challenging data sets with unnatural postures, which occur far more often than expected, the model adaptation algorithm succeeds. A first evaluation showed an average directed Hausdorff distance of 13.22mm and an average DICE overlap of 0.62. Furthermore, we review some of the challenges we encountered during the course of building our model from image data, taken from actual radiation therapy planing cases.

Show publication details

Wichmann, Julian L.; Booz, Christian; Wesarg, Stefan; Kafchitsas, Konstantinos; Bauer, Ralf W.; Kerl, J. Matthias; Lehnert, Thomas; Vogl, Thomas J.; Khan, M. Fawad

Dual-Energy CT-based Phantomless in Vivo Three-dimensional Bone Mineral Density Assessment of the Lumbar Spine

2014

Radiology, Vol.271 (2014), 3, pp. 778-784

Purpose: To evaluate the feasibility of phantomless in vivo dual-energy computed tomography (CT)-based three-dimensional (3D) bone mineral density (BMD) assessment in comparison with dual x-ray absorptiometry (DXA). Materials and Methods: This retrospective study was approved by the institutional review board, and the requirement to obtain informed consent was waived. Data from clinically indicated dual-energy CT and DXA examinations within 2 months, comprising the lumbar spine of 40 patients, were included. By using automated dedicated postprocessing dual-energy CT software, the trabecular bone of lumbar vertebrae L1-L4 were analyzed and segmented. A mixed-effects model was used to assess the correlations between BMD values derived from dual-energy CT and DXA. Results: One hundred sixty lumbar vertebrae were analyzed in 40 patients (mean age, 57.1 years; range, 24-85 years), 21 male (mean age, 54.3 years; range, 24-85 years) and 19 female (mean age, 58.5 years; range, 31-80 years). Mean BMD of L1-L4 determined with DXA was 0.995 g/ cm², and 18 patients (45%) showed an osteoporotic BMD (T score less than 22.5) of at least two vertebrae. Mean dual-energy CT-based BMD of L1-L4 was 0.254 g/cm³. Bland-Altman analysis with mixed effects demonstrated a lack of correlation between dual-energy CT-based and DXA-based BMD values, with a mean difference of 0.7441 and 95% limits of agreement of 0.7441 +/- 0.4080. Conclusion: Dedicated postprocessing of dual-energy CT data allows for phantomless in vivo BMD assessment of the trabecular bone of lumbar vertebrae and enables freely rotatable color-coded 3D visualization of intravertebral BMD distribution.

Show publication details

Noll, Matthias; Puhl, Julian; Wesarg, Stefan

Enhanced Shadow Detection for 3D Ultrasound

2014

Deserno, Thomas Martin (Ed.) et al.: Bildverarbeitung für die Medizin 2014 : Algorithmen, Systeme, Anwendungen. Proceedings des Workshops. Berlin, Heidelberg, New York: Springer, 2014. (Informatik aktuell), pp. 234-239

Workshop Bildverarbeitung für die Medizin (BVM) <17, 2014, Aachen, Germany>

Ultrasound imaging offers a fast, convenient and save instrument to conduct patient examinations for various medical scenarios. However, depending on the target region, occluding bone segments and other materials cause large and undesirable shadowing artifacts. Thus, the experience of the ultrasound operator is crucial for obtaining an ultrasound without artifacts for patient diagnosis. Even more so, when applying automated image processing algorithms. We therefore like to introduce our automatic 3D ultrasound shadow detection method that employs scan line energy and local image entropy information. Applying the method can help to prevent low quality image acquisitions with large shadowing artifacts by indicating shadow occurrences.

Show publication details

Wesarg, Stefan; Wichmann, Julian L.; Booz, Christian; Erdt, Marius; Kafchitsas, Konstantinos; Khan, M. Fawad

First Clinical Experience with BMD Assessment in Vertebrae Using Dual-Energy CT

2014

Erdt, Marius (Ed.) et al.: Clinical Image-Based Procedures. Translational Research in Medical Imaging : Second International Workshop, CLIP 2013. Held in Conjunction with MICCAI 2013. Berlin, Heidelberg, New York: Springer, 2014. (Lecture Notes in Computer Science (LNCS) 8361), pp. 151-159

International Workshop on Clinical Image-based Procedures (CLIP) <2, 2013, Nagoya, Japan>

Dual-energy CT (DECT) can be performed with state-of-the- art dual-source CT (DSCT) scanners and allows for assessing bone mineral density (BMD). In this work, we present first clinical experience with in vivo BMD assessment of vertebrae based on DECT data which has been acquired with a state-of-the-art DSCT scanner in the clinical routine. In contrast to previous work where we did in vitro tests of our method, we apply it her for the first time to in vivo data and prove the feasibility of our technique in a clinical setting. For 25 patients, DXA as well as DECT data have been acquired and BMD of vertebrae was assessed. Advantages of DECT are its 3D capabilities allowing to compute the spatial BMD distribution and to focus the examination on the trabecular bone. Correlations between both imaging techniques regarding the averaged BMD values per vertebra are only moderate.

Show publication details

Steger, Teena; Drechsler, Klaus; Wesarg, Stefan

Image-Based Bronchoscopy Navigation System Based on CT and C-arm Fluoroscopy

2014

Erdt, Marius (Ed.) et al.: Clinical Image-Based Procedures. Translational Research in Medical Imaging : Second International Workshop, CLIP 2013. Held in Conjunction with MICCAI 2013. Berlin, Heidelberg, New York: Springer, 2014. (Lecture Notes in Computer Science (LNCS) 8361), pp. 15-22

International Workshop on Clinical Image-based Procedures (CLIP) <2, 2013, Nagoya, Japan>

Lung cancer diagnosis requires biopsy of airway tissue, which is mostly done by bronchoscopy. Although preoperative CT is available, intraoperatively only 2D information provided by the bronchoscopic camera and fluoroscopy is used. But, guidance of the bronchoscope to the target site would highly benefit from knowing the exact 3D position of the instrument inside the airways. In this paper, we present a system for preoperative planning and intraoperative navigation during bronchoscopy. The preoperative components are automatic bronchial tree segmentation and skeletonization, semi-automatic tumor segmentation and a virtual fly-through simulation for planning purposes. During the intervention, we apply C-arm pose estimation using a marker plate on the patient table to align preoperative CT and intraoperative fluoroscopy. Thus, we can calculate the current 3D position of the bronchoscope inside the bronchial tree. Evaluation of the system components on patient CT and phantom fluoroscopy images showed promising results with high accuracy and robustness.

Show publication details

Drechsler, Klaus; Knaub, Anton; Oyarzun Laura, Cristina; Wesarg, Stefan

Liver Segmentation in Contrast Enhanced MR Datasets Using a Probabilistic Active Shape and Appearance Model

2014

IEEE Computer Society: IEEE 27th International Symposium on Computer-Based Medical Systems : CBMS 2014. Los Alamitos, Calif.: IEEE Computer Society Conference Publishing Services (CPS), 2014, pp. 523-524

IEEE Symposium on Computer-Based Medical Systems (CBMS) <27, 2014, New York, NY, USA>

The current standard for diagnosing liver tumors is contrast-enhanced multiphase computed tomography. On this basis, several software tools have been developed by different research groups worldwide to support physicians for example in measuring remnant liver volume, analyzing tumors, and planning resections. Several algorithms have been developed to perform these tasks. Most of the time, the segmentation of the liver is at the beginning of the processing chain. Therefore, a vast amount of CT-based liver segmentation algorithms have been developed. However, clinics slowly move from CT as the current gold standard for diagnosing liver diseases towards magnetic resonance imaging. In this work, we utilize a Probabilistic Active Shape Model with an MR specific preprocessing and appearance model to segment the liver in contrast enhanced MR images. Evaluation is based on 8 clinical datasets.

Show publication details

Hammon, Matthias; Cavallaro, Alexander; Erdt, Marius; Dankerl, Peter; Kirschner, Matthias; Drechsler, Klaus; Wesarg, Stefan; Uder, Michael; Janka, Rolf

Model-Based Pancreas Segmentation in Portal Venous Phase Contrast-Enhanced CT Images

2014

Journal of Digital Imaging, Vol.26 (2014), 6, pp. 1082-1090. First published online 08 March 2013

This study aims to automatically detect and segment the pancreas in portal venous phase contrast-enhanced computed tomography (CT) images. The institutional review board of the University of Erlangen-Nuremberg approved this study and waived the need for informed consent. Discriminative learning is used to build a pancreas tissue classifier incorporating spatial relationships between the pancreas and surrounding organs and vessels. Furthermore, discrete cosine and wavelet transforms are used to build texture features to describe local tissue appearance. Classification is used to guide a constrained statistical shape model to fit the data. The algorithm to detect and segment the pancreas was evaluated on 40 consecutive CT data that were acquired in the portal venous contrast agent phase. Manual segmentation of the pancreas was carried out by experienced radiologists and served as reference standard. Threefold cross validation was performed. The algorithm-based detection and segmentation yielded an average surface distance of 1.7 mm and an average overlap of 61.2 % compared with the reference standard. The overall runtime of the system was 20.4 min. The presented novel approach enables automatic pancreas segmentation in portal venous phase contrast-enhanced CT images which are included in almost every clinical routine abdominal CT examination. Reliable pancreatic segmentation is crucial for computer-aided detection systems and an organ-specific decision support.

Show publication details

Becker, Meike; Hansen, Stefan; Wesarg, Stefan; Sakas, Georgios

Path Planning for Multi-port Lateral Skull Base Surgery Based on First Clinical Experiences

2014

Erdt, Marius (Ed.) et al.: Clinical Image-Based Procedures. Translational Research in Medical Imaging : Second International Workshop, CLIP 2013. Held in Conjunction with MICCAI 2013. Berlin, Heidelberg, New York: Springer, 2014. (Lecture Notes in Computer Science (LNCS) 8361), pp. 23-30

International Workshop on Clinical Image-based Procedures (CLIP) <2, 2013, Nagoya, Japan>

Our research project investigates a multi-port minimally-traumatic approach for lateral skull base surgery, where the surgical target shall be reached through up to three drill canals. For this purpose, an accurate path planning is crucial. In the present work, we propose a semiautomatic path planning approach for multi-port minimally-traumatic lateral skull base surgery. The best path combinations are automatically determined by optimizing the angles and distance buffers of the drill canals. We compare the automatically computed path combinations for 20 data sets to those selected manually by two different clinicians. The experiments prove that we can adequately reproduce the clinicians' choice.

Show publication details

Steger, Sebastian; Jung, Florian; Wesarg, Stefan

Personalized Articulated Atlas with a Dynamic Adaptation Strategy for Bone Segmentation in CT- or CT/MR Head & Neck Images

2014

Ourselin, Sébastien (Ed.) et al.: Medical Imaging 2014: Image Processing : Progress in Biomedical Optics and Imaging Vol. 15 No. 35. Bellingham: SPIE Press, 2014. (Proceedings of SPIE 9034), 6 p.

SPIE Medical Imaging Symposium <2014, San Diego, CA, USA>

This paper presents a novel segmentation method for the joint segmentation of individual bones in CT- or CT/MR- head & neck images. It is based on an articulated atlas for CT images that learned the shape and appearance of the individual bones along with the articulation between them from annotated training instances. First, a novel dynamic adaptation strategy for the atlas is presented in order to increase the rate of successful adaptations. Then, if a corresponding CT image is available the atlas can be enriched with personalized information about shape, appearance and size of the individual bones from that image. Using mutual information, this personalized atlas is adapted to an MR image in order to propagate segmentations. For evaluation, a head & neck bone atlas created from 15 manually annotated training images was adapted to 58 clinically acquired head & neck CT datasets. Visual inspection showed that the automatic dynamic adaptation strategy was successful for all bones in 86% of the cases. This is a 22% improvement compared to the traditional gradient descent based approach. In leave-one-out cross validation manner the average surface distance of the correctly adapted items was found to be 0.68 mm. In 20 cases corresponding CT/MR image pairs were available and the atlas could be personalized and adapted to the MR image. This was successful in 19 cases.

Show publication details

Steger, Teena; Wesarg, Stefan

Quantitative Analysis of Marker Segmentation for C-Arm Pose Based Navigation

2014

Romero, Laura M. Roa (Ed.): XIII Mediterranean Conference on Medical and Biological Engineering and Computing : MEDICON 2013. Berlin, Heidelberg, New York: Springer, 2014. (IFMBE Proceedings 41), pp. 487-490

The Mediterranean Conference on Medical and Biological Engineering and Computing (MEDICON) <13, 2013, Seville, Spain>

Intraoperative C-arm fluoroscopy is used for better instrument guidance during bronchoscopy. Unfortunately, C-arm images do not provide depth information. But, offering 3D instrument localization would enable faster and more accurate guidance of the bronchoscope. Using the C-arm pose, this can be achieved by combining intraoperative fluoroscopy with a preoperative CT. Thus, the 3D position of the bronchoscope tip inside the bronchial tree can be located and visualized. We developed a marker plate for C-arm pose estimation, which is placed on the patient table. The markers are made of steel and appear in two different shapes: spheres and sticks. Detecting the markers is essential for the C-arm pose estimation method. In this work, we present and evaluate two detection methods for detecting the projected markers on the fluoroscopy images. Tests on cadaver images showed very good results regarding robustness and precision: For circles and lines, 80% and 85%, respectively, of all visible markers were detected, whereas only 1% and 3%, respectively, of all detected markers were missegmented.

Show publication details

Wichmann, Julian L.; Majenka, Pavel; Beeres, Martin; Kromen, Wolfgang; Schulz, Boris; Wesarg, Stefan; Bauer, Ralf W.; Kerl, J. Matthias; Gruber-Rouh, Tatjana; Hammerstingl, Renate; Vogl, Thomas J.; Lehnert, Thomas

Single-portal-phase Low-tube-voltage Dual-energy CT for Short-term Follow-up of Acute Pancreatitis: Evaluation of CT Severity Index, Interobserver Agreement and Radiation Dose

2014

European Radiology, Vol.24 (2014), 11, pp. 2927-2935

Objectives: To intra-individually compare single-portal-phase low-tube-voltage (100-kVp) computed tomography (CT) with 120-kVp images for short-term follow-up assessment of CT severity index (CTSI) of acute pancreatitis, interobserver agreement and radiation dose. Methods: We retrospectively analysed 66 patients with acute pancreatitis who underwent initial dual-contrast-phase CT (unenhanced, arterial, portal phase) at admission and short-term (mean interval 11.4 days) follow-up dual-contrast-phase dual-energy CT. The 100-kVp and linearly blended images representing 120-kVp acquisition follow-up CT images were independently evaluated by three radiologists using a modified CTSI assessing pancreatic inflammation, necrosis and extrapancreatic complications. Scores were compared with paired t test and interobserver agreement was evaluated using intraclass correlation coefficients (ICC). Results: Mean CTSI scores on unenhanced, portal- and dual-contrast-phase images were 4.9, 6.1 and 6.2 (120 kVp) and 5.0, 6.0 and 6.1 (100 kVp), respectively. Contrast-enhanced series showed a higher CTSI compared to unenhanced images (P<0.05) but no significant differences between single- and dual-contrast-phase series (P>0.7). CTSI scores were comparable for 100-kVp and 120-kVp images (P>0.05). Interobserver agreement was substantial for all evaluated series and subcategories (ICC 0.67-0.93). DLP of single-portal-phase 100-kVp images was reduced by 41 % compared to 120-kVp images (363.8 versus 615.9 mGy cm). Conclusions: Low-tube-voltage single-phase 100-kVp CT provides sufficient information for follow-up evaluation of acute pancreatitis and significantly reduces radiation exposure.

Show publication details

Steger, Sebastian; El Hakimi, Wissam; Wesarg, Stefan

Tracking von Instrumenten auf fluoroskopischen Aufnahmen für die navigierte Bronchoskopie

2014

Deserno, Thomas Martin (Ed.) et al.: Bildverarbeitung für die Medizin 2014 : Algorithmen, Systeme, Anwendungen. Proceedings des Workshops. Berlin, Heidelberg, New York: Springer, 2014. (Informatik aktuell), pp. 228-233

Workshop Bildverarbeitung für die Medizin (BVM) <17, 2014, Aachen, Germany>

Intraoperative C-Bogen-Fluoroskopie dient bei der bronchoskopischen Biopsie zur Lokalisation des Bronchoskops und der Biospiezange innerhalb des Patiententhorax. Bei bekannter C-Bogen Pose ist es möglich, aus der 2D-Position der Instrumentenspitze auf der Fluoroskopie deren 3D-Position innerhalb des Bronchialbaums zu berechnen. Während die Pose mit Hilfe einer Markerplatte auf dem Patiententisch bestimmt werden kann, fehlt bisher eine automatische Verfolgung der Instrumentenspitze auf der kontinuierlichen Fluoroskopie. In dieser Arbeit wird eine solche Tracking-Methode vorgestellt und evaluiert. Erste Experimente an einem Bronchialbaum-Phantom lieferten sehr robuste und präzise Ergebnisse und auch die Echtzeitfähigkeit konnte gezeigt werden.

Show publication details

Oyarzun Laura, Cristina; Drechsler, Klaus; Wesarg, Stefan

Two-Step FEM-Based Liver-CT Registration: Improving Internal and External Accuracy

2014

Ourselin, Sébastien (Ed.) et al.: Medical Imaging 2014: Image Processing : Progress in Biomedical Optics and Imaging Vol. 15 No. 35. Bellingham: SPIE Press, 2014. (Proceedings of SPIE 9034), 7 p.

SPIE Medical Imaging Symposium <2014, San Diego, CA, USA>

To know the exact location of the internal structures of the organs, especially the vasculature, is of great importance for the clinicians. This information allows them to know which structures/vessels will be affected by certain therapy and therefore to better treat the patients. However the use of internal structures for registration is often disregarded especially in physical based registration methods. In this paper we propose an algorithm that uses finite element methods to carry out a registration of liver volumes that will not only have accuracy in the boundaries of the organ but also in the interior. Therefore a graph matching algorithm is used to find correspondences between the vessel trees of the two livers to be registered. In addition to this an adaptive volumetric mesh is generated that contains nodes in the locations in which correspondences were found. The displacements derived from those correspondences are the input for the initial deformation of the model. The first deformation brings the internal structures to their final deformed positions and the surfaces close to it. Finally, thin plate splines are used to refine the solution at the boundaries of the organ achieving an improvement in the accuracy of 71%. The algorithm has been evaluated in CT clinical images of the abdomen.

Show publication details

Gollmer, Sebastian T.; Kirschner, Matthias; Buzug, Thorsten M.; Wesarg, Stefan

Using Image Segmentation for Evaluating 3D Statistical Shape Models Built With Groupwise Correspondence Optimization

2014

Computer Vision and Image Understanding, Vol.125 (2014), pp. 283-303

Statistical shape models (SSMs) are a well-established tool in medical image analysis. The most challenging part of SSM construction, which cannot be solved trivially in 3D, is the establishment of corresponding points, so-called landmarks. A popular approach for solving the correspondence problem is to minimize a groupwise objective function using the optimization by re-parameterization approach. To this end, several objective functions, optimization strategies and re-parameterization functions have been proposed. While previous evaluation studies focused mainly on the objective function, we provide a detailed evaluation of different correspondence methods, objective functions, re-parameterization, and optimization strategies. Moreover and contrary to previous works, we use distance measures that compare landmark shape vectors to the original input shapes, thus adequately accounting for correspondences which undersample certain regions of the input shapes. Additionally, we segment binary expert segmentations to benchmark SSMs constructed from different correspondences. This new evaluation technique overcomes limitations of the correspondence based evaluation and allows for directly quantifying the influence of the correspondence on the expected segmentation accuracy. From our evaluation results we identify pitfalls of the current approach and derive practical recommendations for implementing a groupwise optimization pipeline.

Show publication details

Drechsler, Klaus; Meixner, Steven; Oyarzun Laura, Cristina; Wesarg, Stefan

A Framework for Validation of Vessel Segmentation Algorithms

2013

Rodrigues, Pedro Pereira (Ed.) et al.: Proceedings of CBMS 2013 : The 26th IEEE International Symposium on Computer-Based Medical Systems. New York: IEEE, Inc., 2013, pp. 518-519

IEEE Symposium on Computer-Based Medical Systems (CBMS) <26, 2013, Porto, Portugal>

Validation methods used in literature to evaluate vessel segmentation algorithms suffer to a great extent from objectiveness, reliability and reproducibility. This is because almost each group has its own way to evaluate an algorithms. In this paper, an extendable standardized evaluation framework for quantitative validation of vessel segmentation algorithms is presented. As ground-truth, it uses a physical vascular model to simulate the growth of vessels within organ masks extracted from clinical CT datasets. A set of image- and graph- based evaluation metrics are calculated to analyze various aspects of the algorithms under study. Using the proposed framework helps to meet the aforementioned quality criteria.

Show publication details

Jung, Florian; Kirschner, Matthias; Wesarg, Stefan

A Generic Approach to Organ Detection Using 3D Haar-Like Features

2013

Meinzer, Hans-Peter (Ed.) et al.: Bildverarbeitung für die Medizin 2013 : Algorithmen, Systeme, Anwendungen. Proceedings des Workshops. Berlin; Heidelberg; New York: Springer, 2013. (Informatik aktuell), pp. 320-325

Workshop Bildverarbeitung für die Medizin (BVM) <16, 2013, Heidelberg, Germany>

Automatic segmentation of medical images requires accurate detection of the desired organ as a first step. In contrast to application specific approaches, learning-based object detection algorithms are easily adaptable to new applications. We present a learning-based object detection approach based on the Viola-Jones algorithm. We propose several extensions to the original approach, including a new 3D feature type and a multi-organ detection scheme. The algorithm is used to detect six different organs in CT scans as well as the prostate in MRI data. Our evaluation shows that the algorithm provides fast and reliable detection results in all cases.

Show publication details

El Hakimi, Wissam; Wesarg, Stefan

Accurate Super-Resolution Reconstruction for CT and MR Images

2013

Rodrigues, Pedro Pereira (Ed.) et al.: Proceedings of CBMS 2013 : The 26th IEEE International Symposium on Computer-Based Medical Systems. New York: IEEE, Inc., 2013, pp. 445-448

IEEE Symposium on Computer-Based Medical Systems (CBMS) <26, 2013, Porto, Portugal>

The resolution and accuracy of medical images play an important role for early medical diagnosis, since a wrong resolution may increase the risk of making a poor decision. In practice, magnetic resonance and computed tomography images often suffer from anisotropic resolution, so that the image quality is high only within the slices. In this paper we propose a further development of a previously presented super-resolution approach, to reconstruct isotropic high resolution images from only two orthogonal low resolution data sets. Thereby, voxel uncertainties, which arise during image acquisition and preprocessing, are considered. Furthermore, an adapted inpainting method is introduced to ensure a better initial estimation of missing data. Reconstruction quality is also improved, by combining regional and local information. Experiments on synthetic and clinical data sets reveal significant improvement of image quality and accuracy, yielding better results when compared with conventional reconstruction approaches.

Show publication details

Keyhani, Mohammad Hossein; El Hakimi, Wissam; Wesarg, Stefan

Anisotropy Correction of Medical Image Data Employing Patch Similarity

2013

Rodrigues, Pedro Pereira (Ed.) et al.: Proceedings of CBMS 2013 : The 26th IEEE International Symposium on Computer-Based Medical Systems. New York: IEEE, Inc., 2013, pp. 385-388

IEEE Symposium on Computer-Based Medical Systems (CBMS) <26, 2013, Porto, Portugal>

CT or MR image data is typically anisotropic. But, it is desirable to base image processing as well as diagnosis on isotropic image data. In this work, we propose a novel method for correcting anisotropy of 3D image data sets by employing the recurrence of small 2D patches across different scales. We base our method on previous work dealing with super-resolution of single natural 2D images, show the applicability of that approach also to medical images, and extend it to a 3D solution for anisotropy correction. Our results show that the image quality can be significantly improved. For clinical CT and MRI data, we present feedback from the clinical end user.

Show publication details

Steger, Sebastian; Bozoglu, Y. Nazli; Kuijper, Arjan; Wesarg, Stefan

Application of Radial Ray Based Segmentation to Cervical Lymph Nodes in CT Images

2013

IEEE Transactions on Medical Imagaing, Vol.32 (2013), 5, pp. 888-900

The 3D-segmentation of lymph nodes in CT images is required for staging and disease progression monitoring. Major challenges are shape and size variance, as well as low contrast, image noise, and pathologies. In this paper, radial ray based segmentation is applied to lymph nodes: From a seed point, rays are cast into all directions and an optimization technique determines a radius for each ray based on image appearance and shape knowledge. Lymph node specific appearance cost functions are introduced and their optimal parameters are determined. For the first time, the resulting segmentation accuracy of different appearance cost functions and optimization strategies are compared. Further contributions are extensions to reduce the dependency on the seed point, to support a larger variety of shapes, and to enable interaction. The best results are obtained using graph-cut on a combination of the direction weighted image gradient and accumulated intensities outside a predefined intensity range. Evaluation on 100 lymph nodes shows that with an average symmetric surface distance of 0.41 mm the segmentation accuracy is close to manual segmentation and outperforms existing radial ray and model based methods. The method's inter-observer-variability of 5.9% for volume assessment is lower than the 15.9% obtained using manual segmentation.

Show publication details

Drechsler, Klaus; Erdt, Marius; Linguraru, Marius George; Oyarzun Laura, Cristina; Sharma, Karun; Shekhar, Raj; Wesarg, Stefan

Clinical Image-Based Procedures. From Planning to Intervention: International Workshop, CLIP 2012. Held in Conjunction with MICCAI 2012

2013

Berlin, Heidelberg, New York : Springer, 2013

International Workshop on Clinical Image-based Procedures (CLIP) <1, 2012, Nice, France>

Lecture Notes in Computer Science (LNCS) 7761

CLIP 2012 provided a successful forum for the dissemination of emerging image-based clinical techniques. Specific topics included pre-interventional image segmentation and classification (to support diagnosis and clinical decision making), interventional and surgical planning and analysis of dynamic images, and evaluation, visualization and correction techniques for image-based procedures. Clinical applications covered the skull and the brain, cardiac defects, blood vessels, abdominal organs, and cancer in adults and children. The presentations and discussions around the meeting emphasized current challenges and emerging techniques in image-based procedures, strategies for clinical translation of image-based techniques, the role of computational anatomy and image analysis for surgical planning and interventions, and the contribution of medical image analysis to open and minimally invasive surgery.

Show publication details

Hoßbach, Martin; Gregori, Johannes; Wesarg, Stefan; Günther, Matthias

Design and Analysis of a Calibration-method for Stereo-optical Motion Tracking in MRI using a Virtual Calibration Phantom

2013

Nishikawa, Robert M. (Ed.) et al.: Medical Imaging 2013: Physics of Medical Imaging. Part One : Progress in Biomedical Optics and Imaging. Proceedings Issue. Vol. 14, No. 35. Bellingham: SPIE Press, 2013. (Proceedings of SPIE 8668), pp. 86682E-1 - 86682E-7

SPIE Medical Imaging Symposium <2013, Lake Buena Vista, FL, USA>

Motion tracking for head motion compensation in MRI has been a research topic for several years. However, literature is not giving much attention to the calibration of such setups. We present a method to calibrate the coordinate systems of a stereo-optical camera setup mounted to the MRI head coil. Though using a simple setup and visible instead of infrared light for tracking, it is possible to achieve a sub-millimeter tracking precision. Blue water-filled spheres are positioned throughout the whole MRI imaging volume and detected in images of the tracking cameras as well as MRI scans. In order to register the coordinate systems of both camera system and MRI scanner, a heuristic-enhanced brute-force approach is used to match detected spheres in the different images. Then, a rigid transformation is calculated and applied to the cameras' external parameters to align the coordinate systems. The precision of our setup was evaluated using leave-one-out cross validation both for the camera calibration and the scanner coordinate system registration. We found that the cameras' locations and orientations are correct within 0:03mm and 0:03°, using a number of 45 spheres. Evaluation of the MRI coordinate system registration showed an average reprojection error of 1:1 mm. Influence of a feature point jitter of 0:5 px is 0:03mm for a point close to the cameras and 0:3mm for a point close to the back of the patient's head. Tracked poses are correct within 0:17mm and 0:001°.

Show publication details

Hoßbach, Martin; Gregori, Johannes; Wesarg, Stefan; Günther, Matthias

Head Motion Compensation for Arterial Spin Labeling Using Optical Motion Tracking

2013

Drechsler, Klaus (Ed.) et al.: Clinical Image-Based Procedures. From Planning to Intervention : International Workshop, CLIP 2012. Held in Conjunction with MICCAI 2012. Berlin, Heidelberg, New York: Springer, 2013. (Lecture Notes in Computer Science (LNCS) 7761), pp. 1-8

International Workshop on Clinical Image-based Procedures (CLIP) <1, 2012, Nice, France>

Monitoring of Brain perfusion using Arterial Spin Labeling (ASL) during thrombolysis is an example of an MR procedure that will take over one hour. During this time, patient head motion is inevitable. Among the solutions to this problem is motion tracking, which is used to provide the patient's head motion to the MRI scanner to compensate for this motion. We present an approach for head motion tracking which is designed to fit into clinical workflow: Blue paper dots are attached to the patient's forehead and tracked using a stereo camera setup which is mounted to the MRI head coil and easily removed and replaced. The core algorithms are explained and evaluated. Results from clinical experiments conclude the evaluation.

Show publication details

Drechsler, Klaus; Oyarzun Laura, Cristina; Wesarg, Stefan

Hepatic Vein Segmentation Using Wavefront Propagation and Multiscale Vessel Enhancement

2013

Ourselin, Sebastien (Ed.) et al.: Medical Imaging 2013: Image Processing. Part One : Progress in Biomedical Optics and Imaging. Proceedings Issue. Vol. 14, No. 36. Bellingham: SPIE Press, 2013. (Proceedings of SPIE 8669), pp. 86691A-1 - 86691A-7

SPIE Medical Imaging Symposium <2013, Lake Buena Vista, FL, USA>

Modern volumetric imaging techniques such as CT or MRI, aid in the understanding of a patient's anatomy and pathologies. Depending on the medical use case, various anatomical structures are of interest. Blood vessels play an important role in several applications, e.g. surgical planning. Manual delineation of blood vessels in volumetric images is error prone and time consuming. Automated vessel segmentation is a challenging problem due to acquisition-dependent problems such as noise, contrast, spatial resolution, and artifacts. In this paper, a vessel segmentation method is presented that combines a wavefront propagation technique with Hessian-based vessel enhancement. The latter has proven its usefulness as a preprocessing step to detect tubular structures before the actual segmentation is carried out. The former allows for an ordered growing process, which enables topological analysis. The contribution of this work is as follows. 1. A new vessel enhancement filter for tubular structures based on the Laplacian is proposed, 2. a wavefront propagation technique is proposed that prevents leaks by imposing a threshold on the maximum number of voxels that the propagating front must contain, and 3. a volumetric hole filling method is proposed to fill holes, bays, and tunnels which are caused at locations where the tubular structure assumption is violated. The proposed method reduces approximately 50% of the necessary eigenvalue calculations for vessel enhancement and prevents leaks starting at small spots, which usually occur using standard region growing. Qualitative and quantitative evaluation based on several metrics (statistical measures, dice and symmetric average surface distance) is presented.

Show publication details

Steger, Teena; Hoßbach, Martin; Wesarg, Stefan

Marker Detection Evaluation by Phantom and Cadaver Experiments for C-arm Pose Estimation Pattern

2013

Holmes III, David R. (Ed.) et al.: Medical Imaging 2013: Image-Guided Procedures, Robotic Interventions, and Modeling : Progress in Biomedical Optics and Imaging. Proceedings Issue. Vol. 14, No. 38. Bellingham: SPIE Press, 2013. (Proceedings of SPIE 8671), pp. 86711V-1 - 86711V-9

SPIE Medical Imaging Symposium <2013, Lake Buena Vista, FL, USA>

C-arm fluoroscopy is used for guidance during several clinical exams, e.g. in bronchoscopy to locate the bronchoscope inside the airways. Unfortunately, these images provide only 2D information. However, if the C-arm pose is known, it can be used to overlay the intrainterventional fluoroscopy images with 3D visualizations of airways, acquired from preinterventional CT images. Thus, the physician's view is enhanced and localization of the instrument at the correct position inside the bronchial tree is facilitated. We present a novel method for C-arm pose estimation introducing a marker-based pattern, which is placed on the patient table. The steel markers form a pattern, allowing to deduce the C-arm pose by use of the projective invariant cross-ratio. Simulations show that the C-arm pose estimation is reliable and accurate for translations inside an imaging area of 30 cm x 50 cm and rotations up to 30°. Mean error values are 0.33 mm in 3D space and 0.48 px in the 2D imaging plane. First tests on C-arm images resulted in similarly compelling accuracy values and high reliability in an imaging area of 30 cm x 42.5 cm. Even in the presence of interfering structures, tested both with anatomy phantoms and a turkey cadaver, high success rates over 90% and fully satisfying execution times below 4 sec for 1024 px × 1024 px images could be achieved.

Show publication details

Steger, Teena; Steger, Sebastian; Wesarg, Stefan

Marker Removal for C-Arm Pose Estimation Based Bronchoscope Navigation Using Image Inpainting

2013

Österreichische Gesellschaft für Biomedizinische Technik (ÖGBMT): Proceedings BMT (Biomedizinische Technik) 2013 - Dreiländertagung der Deutschen, Schweizerischen und Österreichischen Gesellschaft für Biomedizinische Technik : Biomedizinische Technik - Biomedical Engineering, Vol. 58 (2013) Heft SI-1. Berlin: de Gruyter, 2013, pp. 637-638

Dreiländertagung der Deutschen, Schweizerischen und Österreichischen Gesellschaft für Biomedizinische Technik (BMT) <2013, Graz, Austria>

Using marker-based C-arm pose estimation during bronchoscopy in combination with a preoperative CT segmentation of the bronchial tree, the 3D spatial position of the bronchoscope tip inside the airways can be deter-mined. Naturally, the markers used for pose estimation appear on the fluoroscopy images, which can be disturbing for the physician`s view. To address this issue, we deploy inpainting to remove the reprojected markers from the fluoroscopy. The method was applied to phantom and cadaver images acquired from almost 50 different C-arm poses in total. Pose estimation was successful in 97, 5% of the test cases. Employing our inpainting method to these cases resulted in thorough removal of all markers in the fluoroscopy thereby ostensibly preserving anatomy structures.

Show publication details

Landesberger, Tatiana von; Andrienko, Gennady; Andrienko, Natalia; Bremm, Sebastian; Kirschner, Matthias; Wesarg, Stefan; Kuijper, Arjan

Opening Up the "Black Box" of Medical Image Segmentation with Statistical Shape Models

2013

The Visual Computer, Vol.29 (2013), 9, pp. 893-905

The importance of medical image segmentation increases in fields like treatment planning or computer aided diagnosis. For high quality automatic segmentations, algorithms based on statistical shape models (SSMs) are often used. They segment the image in an iterative way. However, segmentation experts and other users can only asses the final segmentation results, as the segmentation is performed in a "black box manner". Users cannot get deeper knowledge on how the (possibly bad) output was produced. Moreover, they do not see whether the final output is the result of a stabilized process. We present a novel Visual Analytics method, which offers this desired deeper insight into the image segmentation. Our approach combines interactive visualization and automatic data analysis. It allows the expert to assess the quality development (convergence) of the model both on global (full organ) and local (organ areas, landmarks) level. Thereby, local patterns in time and space, e.g., non-converging parts of the organ during the segmentation, can be identified. The localization and specifications of such problems helps the experts creating segmentation algorithms to identify algorithm drawbacks and thus it may point out possible ways how to improve the algorithms systematically. We apply our approach on real-world data showing its usefulness for the analysis of the segmentation process with statistical shape models.

Show publication details

Keil, Matthias; Kaisler, Alexander; Oyarzun Laura, Cristina; Wesarg, Stefan

Refinement and Expansion of Matched Vessel Graphs for Intraoperative Deformable Registration of Hepatic CT and Ultrasound

2013

Holmes III, David R. (Ed.) et al.: Medical Imaging 2013: Image-Guided Procedures, Robotic Interventions, and Modeling : Progress in Biomedical Optics and Imaging. Proceedings Issue. Vol. 14, No. 38. Bellingham: SPIE Press, 2013. (Proceedings of SPIE 8671), pp. 867107-1 - 867107-10

SPIE Medical Imaging Symposium <2013, Lake Buena Vista, FL, USA>

Multimodal registration of intraoperative ultrasound and preoperative contrast enhanced computed tomography (CT) imaging is the basis for image guided percutaneous hepatic interventions. Currently, the surgeon manually performs a rigid registration using vessel structures and other anatomical landmarks for visual guidance. We have previously presented our approach for an automation of this intraoperative registration step based on the definition of bijective correspondences between the vessel structures using an automatic graph matching.¹ This paper describes our method for refinement and expansion of the matched vessel graphs, resulting in a high number of bijective correspondences. Based on these landmarks, we could extend our method to a fully deformable registration. Our system was applied successfully on CT and ultrasound data of nine patients, which are studied in this paper. The number of corresponding vessel points could be raised from a mean of 9.6 points after the graph matching to 70.2 points using the presented refinement method. This allows for the computation of a smooth deformation field. Furthermore, we can show that our deformation calculation raises the registration accuracy for 3 of the 4 chosen target vessels in pre-/postoperative CT with a mean accuracy improvement of 44%.

Show publication details

Erdt, Marius; Knapp, Oliver; Drechsler, Klaus; Wesarg, Stefan

Region Detection in Medical Images Using HOG Classifiers and a Body Landmark Network

2013

Novak, Carol L. (Ed.) et al.: Medical Imaging 2013: Computer-Aided Diagnosis. Part One. Bellingham: SPIE Press, 2013. (Proceedings of SPIE 8670), pp. 867004-1 - 867004-7

SPIE Medical Imaging Symposium <2013, Lake Buena Vista, FL, USA>

Automatic detection of anatomical structures and regions in 3D medical images is important for several computer aided diagnosis tasks. In this work, a new method for simultaneous detection of multiple anatomical areas is proposed. The method consists of two steps: first, single rectangular region candidates are detected independently using 3D variants of Histograms of Oriented Gradients (HOG) features. These features are robust against small changes between regions in rotation and scale which typically occur between different individuals. In a second step, the positions of the detected candidates are refined by incorporating a body landmark network that exploits anatomical relations between different structures. The landmark network consists of a principle component based statistical modeling of the relative positions between the detected regions in training images. The method has been evaluated on thoracic/abdominal CT images of the portal venous phase. In 216 CT images, eight different structures have been trained. Results show an increase in performance using the combination of HOGs and the landmark network in comparison to using independent classifiers without anatomical relations.

Show publication details

Hoßbach, Martin; Noll, Matthias; Wesarg, Stefan

Simplified Stereo-Optical Ultrasound Plane Calibration

2013

Bosch, Johan G. (Ed.) et al.: Medical Imaging 2013: Ultrasonic Imaging, Tomography, and Therapy : Progress in Biomedical Optics and Imaging. Proceedings Issue. Vol. 14, No. 42. Bellingham: SPIE Press, 2013. (Proceedings of SPIE 8675), pp. 86750X-1 - 86750X-7

SPIE Medical Imaging Symposium <2013, Lake Buena Vista, FL, USA>

Image guided therapy is a natural concept and commonly used in medicine. In anesthesia, a common task is the injection of an anesthetic close to a nerve under freehand ultrasound guidance. Several guidance systems exist using electromagnetic tracking of the ultrasound probe as well as the needle, providing the physician with a precise projection of the needle into the ultrasound image. This, however, requires additional expensive devices. We suggest using optical tracking with miniature cameras attached to a 2D ultrasound probe to achieve a higher acceptance among physicians. The purpose of this paper is to present an intuitive method to calibrate freehand ultrasound needle guidance systems employing a rigid stereo camera system. State of the art methods are based on a complex series of error prone coordinate system transformations which makes them susceptible to error accumulation. By reducing the amount of calibration steps to a single calibration procedure we provide a calibration method that is equivalent, yet not prone to error accumulation. It requires a linear calibration object and is validated on three datasets utilizing different calibration objects: a 6mm metal bar and a 1:25mm biopsy needle were used for experiments. Compared to existing calibration methods for freehand ultrasound needle guidance systems, we are able to achieve higher accuracy results while additionally reducing the overall calibration complexity.

Show publication details

Oyarzun Laura, Cristina; Plaza, Pablo Bueno; Drechsler, Klaus; Wesarg, Stefan

Skeleton Based Refinement of Multi-Material Volumetric Meshes

2013

Ourselin, Sebastien (Ed.) et al.: Medical Imaging 2013: Image Processing. Part Two : Progress in Biomedical Optics and Imaging. Proceedings Issue. Vol. 14, No. 36. Bellingham: SPIE Press, 2013. (Proceedings of SPIE 8669), pp. 866936-1 - 866936-6

SPIE Medical Imaging Symposium <2013, Lake Buena Vista, FL, USA>

Accurate multi-material mesh generation is necessary for many applications, e.g. image-guided surgery, in which precision is important. For this application, it is necessary to enhance conventional algorithms with physiological information that adds accuracy to the results. There are several approaches working on the generation of such meshes. However, state of the art approaches show inaccuracies in the areas in which thin structures are, e.g. liver vasculature. These algorithms are not able to detect the vessels in areas in which they are narrow and they assign their elements to wrong materials, e.g., parenchyma. We propose to extend two state of the art algorithms, namely that by Boltcheva et al. and that by Pons et al. and enhance them making use of the skeleton of these structures to solve this problem. By analyzing the mesh generated by the aforementioned algorithms one can find several intersections between the mesh belonging to the vessels and the skeleton, showing that some elements must be mismatched. We evaluate the proposed algorithm in 23 clinical datasets of the liver, in which we previously segmented parenchyma and vessels. For quantitative evaluation, the meshes generated with and without skeleton information are compared. The improvements are shown by means of intersection number, volume and length differences of the vasculature mesh using the different methods. The results show an improvement of 65% for the number of intersections, 4% for the volume and 22% for the length.

Show publication details

Becker, Meike; Gutbell, Ralf; Stenin, Igor; Wesarg, Stefan

Towards Automatic Path Planning for Multi-port Minimally-Traumatic Lateral Skull Base Surgery

2013

Drechsler, Klaus (Ed.) et al.: Clinical Image-Based Procedures. From Planning to Intervention : International Workshop, CLIP 2012. Held in Conjunction with MICCAI 2012. Berlin, Heidelberg, New York: Springer, 2013. (Lecture Notes in Computer Science (LNCS) 7761), pp. 59-66

International Workshop on Clinical Image-based Procedures (CLIP) <1, 2012, Nice, France>

We investigate a multi-port minimally-traumatic approach for lateral skull base surgery. Our long-term goal is to automatically determine the best combination of three paths from the skull surface to the surgical target. Since this multi-port approach is not yet in clinical use, it is difficult to define what the best combination of paths is. Therefore, we present a planning tool which allows to manually selecting a combination of three paths. A clinician used this planning tool to select the best combination of three paths for three target structures in computed tomography data sets of 20 patients. From this choice we derived initial constraints for an automatic planning of the best combination of paths.

Show publication details

Steger, Teena; Wesarg, Stefan

Virtual and Augmented 3D Bronchoscopy Navigation Based on C-arm Fluoroscopy

2013

Paul, Lothar (Ed.) et al.: 3D-NordOst 2013 : 16. Anwendungsbezogener Workshop zur Erfassung, Modellierung, Verarbeitung und Auswertung von 3D-Daten. Berlin: GFaI, 2013, pp. 65-74

Anwendungsbezogener Workshop zur Erfassung, Modellierung, Verarbeitung und Auswertung von 3D-Daten (3D-NordOst) <16, 2013, Berlin, Germany>

Instrument guidance during bronchoscopy is a very intricate task, which demands profound anatomical knowledge and good orientation skills from the physician. Apart from bronchoscopic video, intraoperative C-arm images of the instruments inside the airways are the only auxiliary means available. By tracing the 2D position of the instrument on the fluoroscopy back to its 3D position inside the bronchial tree, an effective navigation support is provided. Given the intraoperative C-arm pose, a preoperative segmentation of the airways and automatic tracking of the 2D position of the instrument on the fluoroscopy, this 3D position of the bronchoscopic instruments can be followed during the whole intervention without manual user interaction. In this paper, we present a bronchoscopy navigation system for virtual preoperative planning and augmented intraoperative guidance. The whole system was successfully tested by inserting a bronchoscope into a bronchial tree model under continuous fluoroscopy. During the procedure, the current 3D position of the bronchoscope's tip could be tracked and visualized inside the bronchial tree segmentation.

Show publication details

Landesberger, Tatiana von; Bremm, Sebastian; Kirschner, Matthias; Wesarg, Stefan; Kuijper, Arjan

Visual Analytics for Model-based Medical Image Segmentation: Opportunities and Challenges

2013

Expert Systems with Applications, Vol.40 (2013), 12, pp. 4934-4943

Segmentation of medical images is a prerequisite in clinical practice. Many segmentation algorithms use statistical shape models. Due to the lack of tools providing prior information on the data, standard models are frequently used. However, they do not necessarily describe the data in an optimal way. Model-based segmentation can be supported by Visual Analytics tools, which give the user a deeper insight into the correspondence between data and model result. Combining both approaches, better models for segmentation of organs in medical images are created. In this work, we identify the main tasks and problems in model-based image segmentation. As a proof of concept, we show that already small visual-interactive extensions can be very beneficial. Based on these results, we present research challenges for Visual Analytics in this area.

Show publication details

Steger, Sebastian; Kirschner, Matthias; Wesarg, Stefan

Articulated Atlas for Segmentation of the Skeleton from Head & Neck CT Datasets

2012

IEEE Engineering in Medicine and Biology Society (EMBS): 2012 IEEE International Symposium on Biomedical Imaging: From Nano to Macro : ISBI 2012. New York: IEEE Press, 2012, pp. 1256-1259

IEEE International Symposium on Biomedical Imaging (ISBI) <9, 2012, Barcelona, Spain>

In this paper a novel articulated atlas for the fully automated segmentation of the skeleton from head & neck CT datasets is presented. An individual atlas describing the shape and appearance is created for each individual bone. Principal Component Analysis is used to learn spatial relations between those atlases resulting in a unified articulated atlas. Transformations are parameterized using the matrix exponential to enable linear combinations required for learning. The adaptation to test images considers appearance, distance to bone structures and the trained articulation space. For evaluation, an atlas created from 10 manually labeled training images has been applied to 46 clinically acquired head & neck CT datasets. Visual inspection showed that in 74% of the cases, the adaptation process was successful. In a second experiment leave-one-out validation was used to quantify the segmentation accuracy. The successfully adapted cases resulted in an average volume overlap error of 30.67 and an average symmetric surface distance of 0.76 mm.

Show publication details

Steger, Sebastian; Wesarg, Stefan

Automated Skeleton Based Multi-modal Deformable Registration of Head&Neck Datasets

2012

Ayache, Nicholas (Ed.) et al.: Medical Image Computing and Computer-Assisted Intervention - MICCAI 2012: Part II : 15th International Conference. Berlin; Heidelberg; New York: Springer, 2012. (Lecture Notes in Computer Science (LNCS) 7511), pp. 66-73

International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI) <15, 2012, Nice, France>

This paper presents a novel skeleton based method for the registration of head&neck datasets. Unlike existing approaches it is fully automated, spatial relation of the bones is considered during their registration and only one of the images must be a CT scan. An articulated atlas is used to jointly obtain a segmentation of the skull, the mandible and the vertebrae C1-Th2 from the CT image. These bones are then successively rigidly registered with the moving image, beginning at the skull, resulting in a rigid transformation for each of the bones. Linear combinations of those transformations describe the deformation in the soft tissue. The weights for the transformations are given by the solution of the Laplace equation. Optionally, the skin surface can be incorporated. The approach is evaluated on 20 CT/MRI pairs of head&neck datasets acquired in clinical routine. Visual inspection shows that the segmentation of the bones was successful in all cases and their successive alignment was successful in 19 cases. Based on manual segmentations of lymph nodes in both modalities, the registration accuracy in the soft tissue was assessed. The mean target registration error of the lymph node centroids was 5.33 ± 2.44 mm when the registration was solely based on the deformation of the skeleton and 5.00 ± 2.38 mm when the skin surface was additionally considered. The method's capture range is sufficient to cope with strongly deformed images and it can be modified to support other parts of the body. The overall registration process typically takes less than 2 minutes.

Show publication details

Kirschner, Matthias; Jung, Florian; Wesarg, Stefan

Automatic Prostate Segmentation in MR Images with a Probabilistic Active Shape Model

2012

Barratt, Dean (Workshop Organizer) et al.: PROMISE12. Proceedings : MICCAI 2012 Grand Challenge on Prostate MR Image Segmentation, pp. 28-35

Prostate MR Image Segmentation Challenge (PROMISE12) <2012, Nice, France>

Segmentation of the prostate gland in Magnetic Resonance (MR) images is an important task for image-guided prostate cancer therapy. The low contrast of the prostate to surrounding tissue in MR images makes automatic segmentation very challenging. In this paper, we propose an automatic approach for robust and accurate prostate segmentation in T2-weighted MR scans. We first employ a boosted prostate detector to locate the prostate in the images, and then use a Probabilistic Active Shape Model for the delineation of its contour. Our approach has been quantitatively evaluated on 50 MR images, on which we achieve a median dice coefficient of 0.85 (IQR: 0.09).

Show publication details

Hoßbach, Martin; Gregori, Johannes; Wesarg, Stefan; Günther, Matthias

Clinical Workflow Optimized Calibration-Method for Stereo-Optical Tracking in MRI

2012

Magnetic Resonance Materials in Physics, Biology and Medicine, Vol.25 (2012), 1 Supplement, p. 580

The Annual Scientific Meeting of the European Society for Magnetic Resonance in Medicine and Biology (ESMRMB) <29, 2012, Lisbon, Portugal>

Though motion tracking for head motion compensation in MRI has been a research topic for some time [1-4], literature is not giving much attention to the calibration of such systems. We present a method to calibrate the coordinate systems of a stereo-optical camera setup mounted to the MRI head coil and show that, though using a simple setup and visible light [1], it is possible to achieve a sub-millimeter tracking precision.

Show publication details

Keil, Matthias; Oyarzun Laura, Cristina; Drechsler, Klaus; Wesarg, Stefan

Combining B-Mode and Color Flow Vessel Segmentation for Registration of Hepatic CT and Ultrasound Volumes

2012

Ropinski, Timo (Ed.) et al.: Eurographics Workshop on Visual Computing for Biology and Medicine : EG VCBM 2012. Goslar: Eurographics Association, 2012, pp. 57-64

Eurographics Workshop on Visual Computing for Biology and Medicine (VCBM) <3, 2012, Norrköping, Sweden>

Multimodal registration of intraoperative ultrasound and preoperative computed tomography imaging is the basis for percutaneous hepatic interventions. Currently, a rigid registration is performed manually by the surgeon using vessel structures and other anatomical landmarks for visual guidance. In this work our approach for intraoperative vessel segmentation from two ultrasound imaging modes, namely B-Mode and color flow mode, is presented. This segmentation is an important step for automation of the intraoperative registration which relies on vessel structures visible in contrast enhanced CT and ultrasound volumes. This paper describes the problems that arise when using B-mode ultrasound for segmentation of vessels and how they can be solved by introducing additional vessel information from color flow imaging. On a total number of 21 patients, our system was applied successfully in 15 cases. For nine randomly chosen patients studied in this paper, our system achieves a 3.45 mm accuracy at points used for registration and 5.01 mm for other landmarks which were not used for the registration process.

Show publication details

El Hakimi, Wissam; Wesarg, Stefan

Confidence Map Based Super-resolution Reconstruction

2012

Haynor, David R. (Ed.) et al.: Medical Imaging 2012: Image Processing. Part One : Progress in Biomedical Optics and Imaging. Proceedings Issue. Vol. 13, No. 30. Bellingham: SPIE Press, 2012. (Proceedings of SPIE 8314), pp. 831440-1 - 831440-10

SPIE Medical Imaging Symposium <2012, San Diego, CA, USA>

Magnetic Resonance Imaging and Computed Tomography usually provide highly anisotropic image data, so that the resolution in the slice-selection direction is poorer than in the in-plane directions. An isotropic high-resolution image can be reconstructed from two orthogonal scans of the same object. While combining the different data sets, all input data are usually equally weighted, without considering the fidelity level of each input information. In this paper we introduce a novel super-resolution method, which considers the fidelity level of each input data by introducing an adaptive confidence map. Experimental results on simulated and real data sets have shown the improved accuracy of reconstructed images, whose resolution approximate the original in-plane resolution in all directions. The quality of the reconstructed high resolution image was improved for noiseless input data sets, and even in the presence of different noise types with a low peak signal to noise ratio.

Show publication details

Erdt, Marius; Steger, Sebastian; Wesarg, Stefan

Deformable Registration of MR Images using a Hierarchical Patch Based Approach With a Normalized Metric Quality Measure

2012

IEEE Engineering in Medicine and Biology Society (EMBS): 2012 IEEE International Symposium on Biomedical Imaging: From Nano to Macro : ISBI 2012. New York: IEEE Press, 2012, pp. 1347-1350

IEEE International Symposium on Biomedical Imaging (ISBI) <9, 2012, Barcelona, Spain>

Magnitude Diffusion Weighted Magnetic Resonance Imaging (DW-MRI) is used to detect lesions in abdominal organs such as the liver. For diagnosis and interventional planning, DW-MRI needs to be registered with Magnetic Resonance (MRI) images in order to provide an anatomical context. DWMRI usually generates images that contain large regions with only little image information. This makes it difficult to elastically register those images to images from other imaging modalities. In this work, we apply a deformable registration approach in order to register magnitude DW-MRI data with MRI images. A normalized quality measure is used to sort out local deformations in low image signal areas that cause non-plausible local deformations. An evaluation on 5 patients shows promising results.

Show publication details

Wesarg, Stefan; Kirschner, Matthias; Becker, Meike; Erdt, Marius; Kafchitsas, Konstantinos; Khan, M. Fawad

Dual-energy CT-based Assessment of the Trabecular Bone in Vertebrae

2012

Methods of Information in Medicine, Vol.51 (2012), 5, pp. 398-405

Background: Osteoporosis can cause severe fractures of bone structures. One important indicator for pathology is a lowered bone mineral density (BMD) - conventionally assessed by dual-energy X-ray absorptiometry (DXA). Dual-energy CT (DECT) - being an alternative that is increasingly used in the clinics - allows the computation of the spatial BMD distribution. Objectives: Using DECT, the trabecular bone of vertebrae is examined. Several analysis methods for revealing the bone density distribution as well as appropriate visualization methods for detecting regions of lowered BMD are needed for computer-assisted diagnosis (CAD) of osteoporosis. The hypothesis that DECT is better suited than DXA for the computation of local BMD is investigated. Methods: Building on a model of the interaction of X-rays with bone tissue, novel methods for assessing the spatial structure of the trabecular bone are presented. CAD of DECT image data is facilitated by segmenting the regions of interest interactively and with an Active Shape Model, respectively. The barycentric space of fractional volumes is introduced as a novel means for analyzing bone constitution. For 29 cadaver specimens, DECT as well as DXA has been examined. BMD values derived from both modalities are compared to local force measurements. In addition, clinical data from two patients who underwent DECT scanning for a different reason is analyzed retrospectively. Results: A novel automated delineation method for vertebrae has been successfully applied to DECT data sets. It is shown that localized BMD measurements based on DECT show a stronger linear correlation (R² = 0.8242, linear regression) to local force measurements than density values derived from DXA (R² = 0.4815). Conclusions: DECT based BMD assessment is a method to extend the usage of increasingly acquired DECT image data. The developed DECT based analysis methods in conjunction with the visualization provide more detailed information for both, the radiologist and the orthopedist, compared to standard DXA based analysis.

Show publication details

Gutbell, Ralf; Becker, Meike; Wesarg, Stefan

Ein Prototyp zur Planung von Bohrpfaden für die minimal-invasive Chirurgie an der Otobasis

2012

Tolxdorff, Thomas (Ed.) et al.: Bildverarbeitung für die Medizin 2012 : Algorithmen, Systeme, Anwendungen. Berlin; Heidelberg; New York: Springer, 2012. (Informatik aktuell), pp. 171-176

Workshop Bildverarbeitung für die Medizin (BVM) <15, 2012, Berlin, Germany>

Bei Operationen an der Otobasis ist es wichtig, umliegende Risikostrukturen wie Gesichtsnerv oder Blutgefäße nicht zu verletzen. Bisher legt der Arzt dazu alle Strukturen frei. Unser Forschungsprojekt untersucht nun einen minimal-invasiven Multi-Port-Ansatz. Im Rahmen dieses Projektes haben wir basierend auf SOFA einen Prototypen zur Planung von Bohrpfaden an der Otobasis entwickelt. Damit kann der Arzt sowohl einen Bohrkanal genauer analysieren als auch sich die Menge aller zulässigen Pfade anzeigen lassen, aus der er schließlich die besten Bohrpfade für die Operation auswählt. Dies ist ein erster Schritt, den Arzt bei der Planung von mehreren Bohrkanälen zu unterstützen.

Show publication details

Rahman, Sami ur; Vateva, Tsvetoslava; Wesarg, Stefan

Enhancing Super-resolution Reconstructed Image Quality in 3D MR Images Using Simulated Annealing

2012

Haynor, David R. (Ed.) et al.: Medical Imaging 2012: Image Processing. Part One : Progress in Biomedical Optics and Imaging. Proceedings Issue. Vol. 13, No. 30. Bellingham: SPIE Press, 2012. (Proceedings of SPIE 8314), pp. 831441-1 - 831441-9

SPIE Medical Imaging Symposium <2012, San Diego, CA, USA>

Super-resolution reconstruction (SRR) algorithms are used for getting high-resolution (HR) data from low-resolution observations. In Maximum a posteriori (MAP) based SRR the observation model is employed for estimating a HR image that best reproduces the two low-resolution input data sets. The parameters of the prior play a significant role in the MAP based SRR. This work concentrates on the investigation of the influence of one such parameter, called temperature, on the reconstructed 3D MR images. The existing approaches on SRR in 3D MR images use a constant value for this parameter. We use a cooling schedule similar to simulated annealing for computing the value of the temperature parameter at each iteration of the SRR. We have used 3D MR cardiac data sets in our experiments and have shown that the iterative computation of the temperature which resembles simulated annealing delivers better results.

Show publication details

Drechsler, Klaus; Oyarzun Laura, Cristina; Wesarg, Stefan

Interventional Planning of Liver Resections: An Overview

2012

IEEE Engineering in Medicine and Biology Society (EMBS): 34th Annual International Conference of the IEEE Engineering in Medicine and Biology Society. Proceedings : EMBC 2012. Stoughton: Printing House, 2012, pp. 3744-3747

Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) <34, 2012, San Diego, CA, USA>

Liver cancer is the third most common type of cancer. Among available treatment options, a surgical resection offers the best prognosis for long-term survival. It is important that such a surgical procedure is carefully prepared. Modern computer technology offers convenient ways to simulate different resection scenarios and help to determine the best treatment for a given case. This paper provides a non-exhaustive overview of existing computer-based systems for interventional planning of liver resections. They are reviewed according to their medical use case, e.g. if they support typical or atypical resections.

Show publication details

Drechsler, Klaus; Erdt, Marius; Oyarzun Laura, Cristina; Wesarg, Stefan

Multiphase Risk Assessment of Atypical Liver Resections

2012

Soda, Paolo (Ed.) et al.: 25th IEEE Symposium on Computer-Based Medical Systems and HealthGrid Conference : The 25th IEEE International Symposium on Computer-Based Medical Systems. New York: IEEE, Inc., 2012, 4 p.

IEEE Symposium on Computer-Based Medical Systems (CBMS) <25, 2012, Rome, Italy>

In this work, we present a system based on open-source toolkits that can utilize multiple phases of a liver CT dataset to plan a surgery before the patient enters the operation room. The contributions are an optimized deformable registration of arterial and venous phases of the liver and a methodology to assess surgical risk utilizing both phases. It is shown by the example of two clinical cases, how our system enables the fusion of complementary information into a 3D representation of the patient anatomy. Processing times of single processing steps on a modern machine are quite low, which allows for an integration into clinical routine.

Show publication details

Kirschner, Matthias; Wesarg, Stefan

Regularisierung lokaler Deformation im probabilistischen Active Shape Model

2012

Tolxdorff, Thomas (Ed.) et al.: Bildverarbeitung für die Medizin 2012 : Algorithmen, Systeme, Anwendungen. Berlin; Heidelberg; New York: Springer, 2012. (Informatik aktuell), pp. 328-333

Workshop Bildverarbeitung für die Medizin (BVM) <15, 2012, Berlin, Germany>

Zur robusten und präzisen Segmentierung von Organen in medizinischen Bilddaten werden oft Varianten des Active Shape Models (ASM) verwendet, die über eine Energieminimierung einen Kompromiss zwischen Bildinformation und Vorwissen über die zu erwartende Organform bestimmen. Im probabilistischen ASM (PASM) wird die Plausibilität einer Form mit Hilfe einer Wahrscheinlichkeitsverteilung bewertet. Da diese lediglich globale, nicht aber lokale Formvariation ausreichend modelliert, kann der PASM ungleichmäßige und damit unplausible Segmentierungskonturen erzeugen. In dieser Arbeit wird der PASM um ein lokales Deformationsmodell erweitert, welches zu glatten Segmentierungskonturen führt. Das lokale Deformationsmodell wird an einem linearen PASM zur Lebersegmentierung und einem nichtlinearen PASM zur Wirbelsegmentierung evaluiert. Die Ergebnisse zeigen, dass die Erweiterung quantitativ wie qualitativ bessere Segmentierungen liefert.

Show publication details

Rahman, Sami ur; Thöne, Clara; Wesarg, Stefan; Völker, Wolfram

Simulation Based Patient-specific Optimal Catheter Selection for Right Coronary Angiography

2012

Holmes III, David R. (Ed.) et al.: Medical Imaging 2012: Image-Guided Procedures, Robotic Interventions, and Modeling : Progress in Biomedical Optics and Imaging. Proceedings Issue. Vol. 13, No. 32. Bellingham: SPIE Press, 2012. (Proceedings of SPIE 8316), pp. 83162S-1 - 83162S-8

SPIE Medical Imaging Symposium <2012, San Diego, CA, USA>

Selecting the best catheter prior to coronary angiography significantly reduces the exposure time to radiation as well as the risk of artery punctures and internal bleeding. In this paper we describe a simulation based technique for selecting an optimal catheter for right coronary angiography using the Simulation Open Framework Architecture (SOFA). We simulate different catheters in a patient-specific arteries model, obtain final placement of different catheters and suggest an optimally placed catheter. The patient-specific arteries model is computed from the patient image data acquired prior to the intervention and the catheters are modeled using Finite Element Method (FEM).

Show publication details

Keyhani, Mohammad Hossein; Wesarg, Stefan (Betreuer)

Super-Resolution from a Single Medical 3D Image Data Set

2012

Darmstadt, TU, Master Thesis, 2012

Super-Resolution Reconstruction is a technique for recovering and reconstructing a higher resolution of a low-resolution image. There exist classical and examplebased super-resolution reconstruction methods, and especially a combination of them leads to reasonable results. In this thesis, we combine a classical and an example-based super-resolution reconstruction method for enhancing the resolution of medical image data sets. Moreover, we present a super-resolution reconstruction method for solving the anisotropy problem in CT and MRI 3D image data sets. Our patch-based method improves the quality of low-resolution structures in anisotropic slices using high-quality structures present in isotropic slices of the data set. Consequently, an isotropic super-resolution image is calculated for each anisotropic slice of the data set. For our approach, no further images with subpixel misalignments are necessary as usually needed by classical superresolution strategies, but the information available in the 3D data set suffice for calculating reasonable super-resolution images. Various experiments with real medical data sets prove the eficiency and correctness of our implementation. We are successful in approximating the super-resolution of medical images reasonably, and improving their quality significantly.

Show publication details

Kirschner, Matthias; Becker, Meike; Wesarg, Stefan

3D Active Shape Model Segmentation with Nonlinear Shape Priors

2011

Fichtinger, Gabor (Ed.) et al.: Medical Image Computing and Computer-Assisted Intervention - MICCAI 2011: Part II : 14th International Conference. Berlin; Heidelberg; New York: Springer, 2011. (Lecture Notes in Computer Science (LNCS) 6892), pp. 492-499

International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI) <14, 2011, Toronto, Canada>

The Active Shape Model (ASM) is a segmentation algorithm which uses a Statistical Shape Model (SSM) to constrain segmentations to 'plausible' shapes. This makes it possible to robustly segment organs with low contrast to adjacent structures. The standard SSM assumes that shapes are Gaussian distributed, which implies that unseen shapes can be expressed by linear combinations of the training shapes. Although this assumption does not always hold true, and several nonlinear SSMs have been proposed in the literature, virtually all applications in medical imaging use the linear SSM. In this work, we investigate 3D ASM segmentation with a nonlinear SSM based on Kernel PCA. We show that a recently published energy minimization approach for constraining shapes with a linear shape model extends to the nonlinear case, and overcomes shortcomings of previously published approaches. Our approach for nonlinear ASM segmentation is applied to vertebra segmentation and evaluated against the linear model.

Show publication details

Kirschner, Matthias; Wesarg, Stefan

Active Shape Models Unleashed

2011

Haynor, David R. (Ed.) et al.: Medical Imaging 2011: Image Processing. Part One : Progress in Biomedical Optics and Imaging. Proceedings Issue. Vol. 13, No. 30. Bellingham: SPIE Press, 2012. (Proceedings of SPIE 8314), pp. 796211-1 - 796211-9

SPIE Medical Imaging Symposium <2011, Lake Buena Vista, FL, USA>

Active Shape Models (ASMs) are a popular family of segmentation algorithms which combine local appearance models for boundary detection with a statistical shape model (SSM). They are especially popular in medical imaging due to their ability for fast and accurate segmentation of anatomical structures even in large and noisy 3D images. A well-known limitation of ASMs is that the shape constraints are over-restrictive, because the segmentations are bounded by the Principal Component Analysis (PCA) subspace learned from the training data. To overcome this limitation, we propose a new energy minimization approach which combines an external image energy with an internal shape model energy. Our shape energy uses the Distance From Feature Space (DFFS) concept to allow deviations from the PCA subspace in a theoretically sound and computationally fast way. In contrast to previous approaches, our model does not rely on post-processing with constrained free-form deformation or additional complex local energy models. In addition to the energy minimization approach, we propose a new method for liver detection, a new method for initializing an SSM and an improved k-Nearest Neighbour (kNN)-classifier for boundary detection. Our ASM is evaluated with leave-one-out tests on a data set with 34 tomographic CT scans of the liver and is compared to an ASM with standard shape constraints. The quantitative results of our experiments show that we achieve higher segmentation accuracy with our energy minimization approach than with standard shape constraints.

Show publication details

Gong, Li; Wesarg, Stefan (Betreuer)

Analysis of the Mitral Valves Based on High Resolution MR Images

2011

Darmstadt, TU, Bachelor Thesis, 2011

Today's technology allows physicians and patients with non-invasive methods to look inside the heart in an easy and fast way. Especially for the heart valves it is practical since they are important for the functionality of the heart and very difficult to analyze without technical support. Currently, mitral valve segmentation has been studied thoroughly in the medical image analysis literature mostly using CT or TEE, but it has not been tried on MR images. This thesis presents a first approach to detect mitral valves on MR images. It provides a semi-automatic segmentation method that is used on MR high resolved datasets. The developed algorithm is based on line tracing which requires initial placement of two seed point respectively on the left and right valve. From the seed point the best next point is then calculated for the next iteration, assuming that the next point has similar grayscale value like the current point. When comparing the neighboring values, thresholding is used which is based on an optimal threshold determination method using the k-means clustering algorithm? In addition, polar grids were introduced to improve the convergence of the algorithm and to distinguish opened and closed mitral valves. Moreover the algorithm can successfully segment the closed mitral valves. The algorithm is also visualized using meshes on the 3D volume. Additional comparisons with manual segmentation made by an expert were done to evaluate the algorithm and put them in comparison.

Show publication details

Becker, Meike; Kirschner, Matthias; Fuhrmann, Simon; Wesarg, Stefan

Automatic Construction of Statistical Shape Models for Vertebrae

2011

Fichtinger, Gabor (Ed.) et al.: Medical Image Computing and Computer-Assisted Intervention - MICCAI 2011: Part II : 14th International Conference. Berlin; Heidelberg; New York: Springer, 2011. (Lecture Notes in Computer Science (LNCS) 6892), pp. 500-507

International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI) <14, 2011, Toronto, Canada>

For segmenting complex structures like vertebrae, a priori knowledge by means of statistical shape models (SSMs) is often incorporated. One of the main challenges using SSMs is the solution of the correspondence problem. In this work we present a generic automated approach for solving the correspondence problem for vertebrae. We determine two closed loops on a reference shape and propagate them consistently to the remaining shapes of the training set. Then every shape is cut along these loops and parameterized to a rectangle. There, we optimize a novel combined energy to establish the correspondences and to reduce the unavoidable area and angle distortion. Finally, we present an adaptive resampling method to achieve a good shape representation. A qualitative and quantitative evaluation shows that using our method we can generate SSMs of higher quality than the ICP approach.

Show publication details

Erdt, Marius; Kirschner, Matthias; Drechsler, Klaus; Wesarg, Stefan; Hammon, Matthias; Cavallaro, Alexander

Automatic Pancreas Segmentation in Contrast Enhanced CT Data Using Learned Spatial Anatomy and Texture Descriptors

2011

IEEE Engineering in Medicine and Biology Society (EMBS): 2011 IEEE International Symposium on Biomedical Imaging: From Nano to Macro : ISBI 2011. New York: IEEE Press, 2011, pp. 2076-2082

IEEE International Symposium on Biomedical Imaging (ISBI) <8, 2011, Chicago, IL, USA>

Pancreas segmentation in 3-D computed tomography (CT) data is of high clinical relevance, but extremely difficult since the pancreas is often not visibly distinguishable from the small bowel. So far no automated approach using only single phase contrast enhancement exist. In this work, a novel fully automated algorithm to extract the pancreas from such CT images is proposed. Discriminative learning is used to build a pancreas tissue classifier that incorporates spatial relationships between the pancreas and surrounding organs and vessels. Furthermore, discrete cosine and wavelet transforms are used to build computationally inexpensive but meaningful texture features in order to describe local tissue appearance. Classification is then used to guide a constrained statistical shape model to fit the data. Cross-validation on 40 CT datasets yielded an average surface distance of 1.7 mm compared to ground truth which shows that automatic pancreas segmentation from single phase contrast enhanced CT is feasible. The method even outperforms automatic solutions using multiple-phase CT both in accuracy and computation time.

Show publication details

Kirschner, Matthias; Wesarg, Stefan

Automatische Initialisierung von Formmodellen mittels modellbasierter Registrierung

2011

Handels, H. (Ed.) et al.: Bildverarbeitung für die Medizin 2011 : Algorithmen, Systeme, Anwendungen. Berlin; Heidelberg; New York: Springer, 2011. (Informatik aktuell), pp. 69-73

Workshop Bildverarbeitung für die Medizin (BVM) <14, 2011, Aachen, Germany>

Das Active Shape Model (ASM) ist ein Segmentierungsverfahren, das statistische Formmodelle (SFM) verwendet, um Organe in Bilddaten trotz geringen Kontrastes zu benachbarten Strukturen robust und effizient zu segmentieren. Da das ASM ein lokales Suchverfahren ist, muss vor der Segmentierung zunächst das gesuchte Organ im Bild detektiert werden, um dann das SFM initial möglichst genau zu platzieren. In dieser Arbeit stellen wir ein neues Verfahren zur Modellinitialisierung vor. Unser Hauptbeitrag ist eine neue Variante des Iterative Closest Point Algorithmus (ICP), die es erlaubt, ein komplettes SFM mit einer Punktmenge effizient zu registrieren. Das Verfahren wird zur Detektion der Leber mit anschließender SFM Initialisierung in 14 kontrastverstärkten CT-Aufnahmen eingesetzt und quantitativ evaluiert.

Show publication details

Wesarg, Stefan; Erdt, Marius; Kafchitsas, Konstantinos; Khan, M. Fawad

Direct Visualization of Regions with Lowered Bone Mineral Density in Dual-Energy CT Images of Vertebrae

2011

Summers, Ronald M. (Ed.) et al.: Medical Imaging 2011: Computer-Aided Diagnosis. Part Two : Progress in Biomedical Optics and Imaging. Proceedings Issue. Vol. 12, No. 32. Bellingham: SPIE Press, 2011. (Proceedings of SPIE 7963), pp. 79633J-1 - 79633J-10

SPIE Medical Imaging Symposium <2011, Lake Buena Vista, FL, USA>

Dual-energy CT allows for a better material differentiation than conventional CT. For the purpose of osteoporosis diagnosis, a detection of regions with lowered bone mineral density (BMD) is of high clinical interest. Based on an existing biophysical model of the trabecular bone in vertebrae a new method for directly highlighting those low density regions in the image data has been developed. For this, we combine image data acquired at 80 kV and 140 kV with information about the BMD range in different vertebrae and derive a method for computing a color enhanced image which clearly indicates low density regions. An evaluation of our method which compares it with a quantitative method for BMD assessment shows a very good correspondence between both methods. The strength of our method lies in its simplicity and speed.

Show publication details

Becker, Meike; Kirschner, Matthias; Wesarg, Stefan

Konsistente Parametrisierung von Flächen vom Geschlecht 1 zur Bildung eines statistischen Formmodells des Wirbels

2011

Handels, H. (Ed.) et al.: Bildverarbeitung für die Medizin 2011 : Algorithmen, Systeme, Anwendungen. Berlin; Heidelberg; New York: Springer, 2011. (Informatik aktuell), pp. 74-78

Workshop Bildverarbeitung für die Medizin (BVM) <14, 2011, Aachen, Germany>

Für die Segmentierung komplexer Strukturen wie beispielsweise Wirbel werden häufig statistische Formmodelle (SFM) verwendet. Bei der Konstruktion des SFM stellt die Lösung des Korrespondenzproblems eine der größten Herausforderungen dar. In dieser Arbeit präsentieren wir einen neuen automatischen Ansatz für die Initiallösung des Korrespondenzproblems für Flächen vom Geschlecht 1. Dazu schneiden wir eine Referenzfläche der Trainingsmenge entlang zweier möglichst kurzer Schleifen auf und propagieren diese auf die übrigen Flächen der Trainingsmenge. Anschließend bilden wir jede Fläche auf den Parameterraum des Rechtecks ab, wo wir die entstehende Flächenverzerrung mit einem heuristischen Ansatz verringern. Damit können wir SFM mit erhöhter Qualität konstruieren.

Show publication details

Wesarg, Stefan

Monitoring Treatment Outcome: A Visualization Prototype for Left Ventricular Transformation

2011

Sachse, Frank B. (Ed.) et al.: Functional Imaging and Modeling of the Heart. Proceedings. Berlin, Heidelberg, New York: Springer, 2011. (Lecture Notes in Computer Science (LNCS) 6666), pp. 121-128

International Conference on Functional Imaging and Modeling of the Heart (FIMH) <6, 2011, New York City, NY, USA>

The analysis of cardiac dynamics - especially of the left ventricle - is a means for evaluating the healthiness of the heart. In case that a malfunction has been detected and afterwards has been treated, the question arises whether the treatment was successful or not. On a longer time scale, it is of clinical interest to compare the results of follow-up studies with those of former examinations. In this paper, we address both issues by presenting a visualization prototype for the comparison of left ventricular dynamics obtained from cine-MRI data. Our approach is based on the computation of differences for standard cardiac parameters between two time series which have been acquired prior to and after treatment. For their visualization, we use a series of bull's-eye displays allowing for an in-depth examination of the treatment outcome. Here, we focus on the special clinical application ventricular reduction surgery where we perform a retrospective evaluation for cine-MRI data acquired prior to and right after surgery as well as several months later. We compare our results with diagnosis information obtained from clinical experts.

Show publication details

Kirschner, Matthias; Gollmer, Sebastian T.; Wesarg, Stefan; Buzug, Thorsten M.

Optimal Initialization for 3D Correspondence Optimization: An Evaluation Study

2011

Székely, Gábor (Ed.) et al.: Information Processing in Medical Imaging : 22nd International Conference, IPMI 2011. Berlin, Heidelberg, New York: Springer, 2011. (Lecture Notes in Computer Science (LNCS) 6801), pp. 308-319

International Conference on Information Processing in Medical Imaging (IPMI) <22, 2011, Kloster Irsee, Germany>

The identification of corresponding landmarks across a set of training shapes is a prerequisite for statistical shape model (SSM) construction. We automatically establish 3D correspondence using one new and several known alternative approaches for consistent, shape-preserving, spherical parameterization. The initial correspondence determined by all employed methods is refined by optimizing a groupwise objective function. The quality of all models before and after optimization is thoroughly evaluated using several data sets of clinically relevant, anatomical objects of varying complexity. Correspondence quality is benchmarked in terms of the SSMs' specificity and generalization ability, which are measured using different surface based distance functions. We find that our new approach performs best for complex objects. Furthermore, all new and previously published methods of our own allow for (i) building SSMs that are significantly better than the well-known SPHARM method, (ii) establishing quasi-optimal correspondence for low and moderately complex objects without additional optimization, and (iii) considerably speeding up convergence, thus, providing means for practical, fast, and accurate SSM construction.

Show publication details

Rahman, Sami ur; Wesarg, Stefan; Völker, Wolfram

Patient Specific Optimal Catheter Selection for Right Coronary Artery

2011

Wong, Kenneth H. (Ed.) et al.: Medical Imaging 2011: Visualization, Image-Guided Procedures, and Modeling. Part One : Progress in Biomedical Optics and Imaging. Proceedings Issue. Vol. 12, No. 33. Bellingham: SPIE Press, 2011. (Proceedings of SPIE 7964), pp. 796412-1 - 796412-7

SPIE Medical Imaging Symposium <2011, Lake Buena Vista, FL, USA>

During coronary artery angiography, a catheter is used to inject a contrast dye into the coronary arteries. Due to the anatomical variation of the aorta and the coronary arteries in different humans, one common catheter cannot be used for all patients. The cardiologists test different catheters for a patient and select the best catheter according to the patient's anatomy. This procedure is time consuming and there is a slight chance of cancer from excessive exposure to radiation. To overcome these problems, we propose a computer aided catheter selection procedure. In this paper we present our approach for the angiography of the Right Coronary Artery (RCA). Our approach involves segmentation of the aorta and coronary arteries, finding the centerline and computing the Curve Angle (CA) and Curve Length (CL) between the aorta and the coronary arteries. We then compute CA and CL of catheters and suggest a catheter with the closest CA and CL with respect to the aorta's and coronary arteries' CA and CL. This solution avoids testing of many catheters during catheterization. The cardiologist already gets the recommendation about the optimal catheter for the patient prior to the intervention.

Show publication details

Rahman, Sami ur; Wesarg, Stefan; Völker, Wolfram

Patient Specific Optimal Catheter Selection for the Left Coronary Artery

2011

Gesellschaft für Informatik (GI): Emerging Technologies for Medical Diagnosis and Therapy : Workshop der 41. Jahrestagung der Gesellschaft für Informatik e.V. (GI): Informatik 2011 - Informatik schafft Communities [online]. [cited 14 December 2011] Available from: http://www.user.tu-berlin.de/komm/CD/html/ws905.html, 2011, 9 p.

Informatik 2011 - Jahrestagung der Gesellschaft für Informatik e.V. (GI) <41, 2011, Berlin, Germany>

During coronary artery angiography, a catheter is used to inject a contrast dye into the coronary arteries. Due to the anatomical variation of the aorta and the coronary arteries in different humans, one common catheter cannot be used for all patients. The cardiologists test different catheters for a patient and select the best one according to the patient's anatomy. To overcome these problems, we propose a computer aided catheter selection procedure. In this paper we present our approach for patient specific optimal catheter selection for the angiography of the left coronary artery (LCA). It involves segmentation of the aorta and coronary arteries, finding the centerline and computing some geometric parameters. These parameters include curve angle of the LCA, LCA contralateral wall curve length, and the aorta cavity length. We then consider catheters for the LCA and compute the angles and lengths of the two curves as well as the distance between these curves. We suggest a catheter that is the closest one to the patient's arteries geometry. This solution avoids testing of many catheters during catheterization. The cardiologist already gets a recommendation about the optimal catheter for the patient prior to the intervention.

Show publication details

Wesarg, Stefan; Becker, Meike; El Hakimi, Wissam

Planning of High Precision Surgery at the Lateral Skull Base

2011

Deutsches Krebsforschungszentrum (DKFZ): Advanced Medical Image Processing. Proceedings 2011. Germany, 2011, pp. 72-78

Korea-Germany Joint Workshop on Advanced Medical Image Processing <14, 2011, Heidelberg, Germany>

Minimally traumatic surgery at the lateral skull base is currently far from being the standard in clinical routine. Drilling one or more linear paths towards the inner ear requires high accuracy. The project MUKNO aims at the development of tools and methods for allowing those minimally invasive approaches. For this, the errors of each single step of the surgery pipeline shall be quantified and minimized. All subsequent steps (e. g. surgery planning, navigation) rely on the acquired image data. Thus, it is crucial to provide such data with a sufficient accuracy. In this work, we present first results obtained in the project MUKNO. Here, one aspect is the high resolution image reconstruction from two orthogonal data sets. A second topic is the integration of a patient specific model of the lateral skull base into a virtual surgery simulation for the determination of possible drilling paths. For this, a first prototype of a planning tool is presented.

Show publication details

Wesarg, Stefan; Erdt, Marius; Kafchitsas, Konstantinos; Khan, M. Fawad

Quantifizierung und Visualisierung der Struktur des trabekulären Knochens in Wirbelkörpern

2011

Handels, H. (Ed.) et al.: Bildverarbeitung für die Medizin 2011 : Algorithmen, Systeme, Anwendungen. Berlin; Heidelberg; New York: Springer, 2011. (Informatik aktuell), pp. 309-313

Workshop Bildverarbeitung für die Medizin (BVM) <14, 2011, Aachen, Germany>

Dual-Energy CT (DECT) ist seit einigen Jahren im klinischen Einsatz. Wir stellen eine Methode vor, wie mittels DECT die räumliche Verteilung der Knochendichte in Wirbeln bestimmt und für die Diagnose von Osteoporose verwendet werden kann. Des Weiteren beschreiben wir einen neuen Ansatz für die Bestimmung der Zusammensetzung des trabekulären Knochens. Unsere Methoden wurden erfolgreich an DECT-Daten von 29 Wirbeln getestet und bedeuten einen Informationsgewinn für den Bereich der Orthopädie.

Show publication details

Lassmann, Andreas; Wesarg, Stefan (Betreuer)

Simulation der Festigkeit von Pedikelschrauben in Wirbeln

2011

Darmstadt, TU, Bachelor Thesis, 2011

Diese Bachelorarbeit behandelt die Simulation der Festigkeit von Pedikelschrauben in Wirbeln. Ich gebe einen Einblick in das medizinische Umfeld und leite daraus ein physikalisches Modell ab. Damit soll geklärt werden, ob sich die Haltekraft der Pedikelschraube anhand der Knochendichte bestimmen lässt. Zur Lösung des Problems wird ein Simulationswerkzeug verwendet, mit dessen Hilfe eine Finite Elemente Analyse durchgeführt werden kann. Die dazu benötigte Finite Elemente Methode wird für die Berechnung der Kraft modifiziert. Zusätzlich wird eine Strategie zum Stoppen der Simulation umgesetzt. Ich zeige mein Vorgehen während der Simulation auf und diskutiere deren Ergebnisse. Bei der Evaluation werden sich Schwankungen und Ausreißer abzeichnen, die im Weiteren diskutiert werden. Zuletzt zeige ich eine alternative Simulationsmethode auf und gebe einen Ausblick auf Verbesserungsmöglichkeiten.

Show publication details

Flehmann, Eugen; Rahman, Sami ur; Wesarg, Stefan; Völker, Wolfram

Towards Patient Specific Catheter Selection: Computation of Aortic Geometry Based on Fused MRI Data

2011

Sachse, Frank B. (Ed.) et al.: Functional Imaging and Modeling of the Heart. Proceedings. Berlin, Heidelberg, New York: Springer, 2011. (Lecture Notes in Computer Science (LNCS) 6666), pp. 145-152

International Conference on Functional Imaging and Modeling of the Heart (FIMH) <6, 2011, New York City, NY, USA>

In coronary angiography, a catheter's tip has to be directed through the aorta towards the ostium - the region where the coronary arteries arise. Due to the anatomical variation in different humans, there is no common catheter which can be used for all patients. Thus, in a trial and error procedure cardiologists find a catheter that fits to the patient's anatomy. To replace this time consuming approach by providing a computer aided planning tool to be used prior to the intervention is the focus of our work. First of all, it is necessary for such a system to derive geometrical parameters for the patient's aorta as well as for the different available catheters. Based thereon, the best fitting catheter can be selected. In this paper, we discuss the first step: the computation of geometrical parameters from the patient's image data. Due to the setting defined by our clinical partner, two MRI data sets are acquired and should be used for the computation. This requires a specific image processing pipeline which we present here and which has to our knowledge not been proposed so far. Furthermore, we show first results obtained for real clinical data sets and discuss the subsequent steps for the development of the catheter selection tool.

Show publication details

Wesarg, Stefan; Kirschner, Matthias; Khan, M. Fawad

2D Histogram Based Volume Visualization: Combining Intensity and Size of Anatomical Structures

2010

International Journal of Computer Assisted Radiology and Surgery, Vol.5 (2010), 6, pp. 655-666. First published as Online First, 30 May 2010. DOI: 10.1007/s11548-010-0480-1

Purpose Surgical planning requires 3D volume visualizations based on transfer functions (TF) that assign optical properties to volumetric image data. Two-dimensional TFs and 2D histograms may be employed to improve overall performance. Methods Anatomical structures were used for 2D TF definition in an algorithm that computes a new structure-size image from the original data set. The original image and structure-size data sets were used to generate a structure-size enhanced (SSE) histogram. Alternatively, the gradient magnitude could be used as second property for 2D TF definition. Both types of 2D TFs were generated and compared using subjective evaluation of anatomic feature conspicuity. Results Experiments with several medical image data sets provided SSE histograms that were judged subjectively to be more intuitive and better discriminated different anatomical structures than gradient magnitude-based 2D histograms. Conclusions In clinical applications, where the size of anatomical structures is more meaningful than gradient magnitude, the 2D TF can be effective for highlighting anatomical structures in 3D visualizations.

Show publication details

Jung, Florian; Wesarg, Stefan

3D Registration Based on Normalized Mutual Information: Performance of CPU vs. GPU Implementation

2010

Deserno, Thomas M. (Ed.) et al.: Bildverarbeitung für die Medizin 2010 : Algorithmen, Systeme, Anwendungen. Berlin; Heidelberg; New York: Springer, 2010. (Informatik aktuell), pp. 325-329

Workshop Bildverarbeitung für die Medizin (BVM) <13, 2010, Aachen, Germany>

Medical image registration is time-consuming but can be sped up employing parallel processing on the GPU. Normalized mutual information (NMI) is a well performing similarity measure for performing multi-modal registration. We present CUDA based solutions for computing NMI on the GPU and compare the results obtained by rigidly registering multi-modal data sets with a CPU based implementation. Our tests with RIRE data sets show a speed-up of factor 5 to 7 for our best GPU implementation.

Show publication details

Kirschner, Matthias; Wesarg, Stefan

3D Statistical Shape Model Building Using Consistent Parameterization

2010

Deserno, Thomas M. (Ed.) et al.: Bildverarbeitung für die Medizin 2010 : Algorithmen, Systeme, Anwendungen. Berlin; Heidelberg; New York: Springer, 2010. (Informatik aktuell), pp. 291-295

Workshop Bildverarbeitung für die Medizin (BVM) <13, 2010, Aachen, Germany>

We propose a new correspondence optimisation algorithm for building 3D statistical shape models (SSMs) of genus-0 shapes. The main contribution of our work is the use of parameter space propagation to generate consistent spherical parameterisations of the training shapes. We present evaluation results for two data sets: A set of 30 liver shapes from different patients, and a set of 25 left ventricles covering the cardiac cycle of a single patient. Our evaluation shows that the use of parameter space propagation improves the robustness of correspondence optimisation algorithms and lead to fast convergence times.

Show publication details

Wesarg, Stefan; Kafchitsas, Konstantinos; Erdt, Marius; Khan, M. Fawad

CAD of Osteoporosis in Vertebrae Using Dual-energy CT

2010

Dillon, Tharam et al.: Twenty-Third IEEE Symposium on Computer-Based Medical Systems : CBMS 2010. Los Alamitos, Calif.: IEEE Computer Society, 2010, pp. 358-363

IEEE Symposium on Computer-Based Medical Systems (CBMS) <23, 2010, Perth, Australia >

The assessment of bone mineral density (BMD) in vertebrae is critical for the diagnosis of osteoporosis. Recent developments in dual-source CT allow for the simultaneous acquisition of two image data sets with different X-ray tube energies - dual-energy CT (DECT). We present a comprehensive approach for assessing the density of the trabecular bone in vertebrae of the spine based on DECT image data. For this, we apply and combine methods from different areas: the deformation of a template mesh for delineating the structures of interest, a biophysical model of the trabecular bone for the computation of BMD values, and different visualization approaches for the display of the results. In addition, we investigate the correlation between the computed BMD values with concurrently measured pull-out forces for pedicle screws. We show that there is a linear correlation between both measures and thus, DECT provides correct BMD values for the trabecular bone. We conclude that our approach enables the radiologist to diagnose osteoporosis based on DECT image data which has the potential to replace the current gold standard dual-energy X-ray absorptiometry.

Show publication details

Rahman, Sami ur; Wesarg, Stefan

Combining Short-axis and Long-axis Cardiac MR Images by Applying a Super-resolution Reconstruction Algorithm

2010

Dawant, Benoit M. (Ed.) et al.: Medical Imaging 2010: Image Processing. Part One : Progress in Biomedical Optics and Imaging. Proceedings Issue. Vol. 11, No. 33. Bellingham: SPIE Press, 2010. (Proceedings of SPIE 7623), pp. 76230I-1-76230I-12

SPIE Medical Imaging Symposium <2010, San Diego, CA, USA>

In cardiac MR images the slice thickness is normally greater than the pixel size within the slices. In general, better segmentation and analysis results can be expected for isotropic high-resolution (HR) data sets. If two orthogonal data sets, e. g. short-axis (SA) and long-axis (LA) volumes are combined, an increase in resolution can be obtained. In this work we employ a super-resolution reconstruction (SRR) algorithm for computing high-resolution data sets from two orthogonal SA and LA volumes. In contrast to a simple averaging of both data in the overlapping region, we apply a maximum a posteriori approach. There, an observation model is employed for estimating an HR image that best reproduces the two low-resolution input data sets. For testing the SRR approach, we use clinical MRI data with an in-plane resolution of 1.5 mm×1.5 mm and a slice thickness of 8 mm. We show that the results obtained with our approach are superior to currently used averaging techniques. Due to the fact that the heart deforms over the cardiac cycle, we investigate further, how the replacement of a rigid registration by a deformable registration as preprocessing step improves the quality of the final HR image data. We conclude that image quality is dramatically enhanced by applying an SRR technique especially for cardiac MR images where the resolution in slice-selection direction is about five times lower than within the slices.

Show publication details

Kirschner, Matthias; Wesarg, Stefan

Construction of Groupwise Consistent Shape Parameterizations by Propagation

2010

Dawant, Benoit M. (Ed.) et al.: Medical Imaging 2010: Image Processing. Part One : Progress in Biomedical Optics and Imaging. Proceedings Issue. Vol. 11, No. 33. Bellingham: SPIE Press, 2010. (Proceedings of SPIE 7623), pp. 762352-1 - 762352-12

SPIE Medical Imaging Symposium <2010, San Diego, CA, USA>

Prior knowledge can highly improve the accuracy of segmentation algorithms for 3D medical images. A popular method for describing the variability of shape of organs are statistical shape models. One of the greatest challenges in statistical shape modeling is to compute a representation of the training shapes as vectors of corresponding landmarks, which is required to train the model. Many algorithms for extracting such landmark vectors work on parameter space representations of the unnormalized training shapes. These algorithms are sensitive to inconsistent parameterizations: If corresponding regions in the training shapes are mapped to different areas of the parameter space, convergence time increases or the algorithms even fail to converge. In order to improve robustness and decrease convergence time, it is crucial that the training shapes are parameterized in a consistent manner. We present a novel algorithm for the construction of groupwise consistent parameterizations for a set of training shapes with genus-0 topology. Our algorithm firstly computes an area-preserving parameterization of a single reference shape, which is then propagated to all other shapes in the training set. As the parameter space propagation is controlled by approximate correspondences derived from a shape alignment algorithm, the resulting parameterizations are consistent. Additionally, the area-preservation property of the reference parameterization is likewise propagated such that all training shapes can be reconstructed from the generated parameterizations with a simple uniform sampling technique. Though our algorithm considers consistency as an additional constraint, it is faster than computing parameterizations for each training shape independently from scratch.

Show publication details

Erdt, Marius; Steger, Sebastian; Kirschner, Matthias; Wesarg, Stefan

Fast Automatic Liver Segmentation Combining Learned Shape Priors with Observed Shape Deviation

2010

Dillon, Tharam et al.: Twenty-Third IEEE Symposium on Computer-Based Medical Systems : CBMS 2010. Los Alamitos, Calif.: IEEE Computer Society, 2010, pp. 249-254

IEEE Symposium on Computer-Based Medical Systems (CBMS) <23, 2010, Perth, Australia >

We present a novel statistical shape model approach for fully automatic CT liver segmentation. Unlike previous techniques, our method combines learned local shape priors with constraints that are directly derived from the current curvature of the model in order to restrict adaptation to regions where large deformations are expected and observed. Our approach is based on a multi-tiered framework that is more robust against model initialization errors than existing methods, because the model's degrees of freedom are step-wise increased. We evaluated our method on a large data base of 86 CT liver scans from different vendors, protocols, varying resolution and contrast enhancement. For comparison, 50 of the scans were taken from 2 public data bases, one of it being the MICCAI'07 liver segmentation challenge data base. Evaluation shows state of the art results with an average mean surface distance between 1.3 mm and 1.85 mm compared to ground truth depending on the image resolution. With an average segmentation time of 45 seconds our approach outperforms other automatic methods.

Show publication details

Ghassemi Hosseini, Azam Sadat; Wesarg, Stefan (Betreuer)

Knochenmineralbestimmung mittels Dual-Energie-Computertomographie

2010

Darmstadt, TU, Diplomarbeit, 2010

Osteoporose ist eine systemische Skeletterkrankung, die durch eine Verminderung der Knochenmasse und Veränderungen in der Knochenmikrostruktur charakterisiert ist. Das in dieser Diplomarbeit vorgestellte Verfahren basiert auf einem Knochendichtemessverfahren, das eine Darstellung der Knochendichte aus entsprechenden dreidimensionalen Dual-Energy-CT-Datensätzen ermöglicht. In einem ersten Schritt wird die spongiose Knochensubstanz in jeder einzelnen CT-Schicht identifiziert und markiert. Anschließend wird die Knochenmineraldichte und die Dichte des trabekulären Raums auf Basis der CT-Zahlen der Datensätze berechnet. Abschließend werden diese Werte den Voxeln der CT-Aufnahmen der Wirbelsäule für eine interaktive Untersuchung der Knochendichte im trabekulären Knochen überlagert.

Show publication details

Erdt, Marius; Schlegel, Patrice; Wesarg, Stefan

Multi-Layer Deformable Models for Medical Image Segmentation

2010

University of Ioannina: 10th International Conference on Information Technology and Applications in Biomedicine. Proceedings CD-ROM : 10th International Conference on Information Technology and Applications in Biomedicine. New York: IEEE, Inc., 2010, 4 p.

International Conference on Information Technology and Applications in Biomedicine (ITAB) <10, 2010, Corfu, Greece>

In this work, a Multi-Layer Deformable Model (MLDM) for medical image segmentation is proposed. In contrast to common deformable model based segmentation approaches our new method incorporates a multi-layer geometric model that allows a sampling of the organ's interior. An adaptation logic processes the additional information gained from interior layers in order to fit the model to the data. The deformation is coupled with a dynamic internal energy function represented by a link-oriented flexibility in order to allow the model to accurately adapt to cavities. Exploiting the additional depth information, our approach detects low contrasted transitions between organs more reliably and recovers better from bad model initialization than existing methods. Our approach has been evaluated using representative CT data sets of the liver as well as CT bladder scans. Evaluation using ground truth data showed that our multi-layer technique yields superior results in contrast to common single surface segmentation. Since the amount of layers is flexible, the most interior regions which only carry little regional information can be excluded from optimization. Together with the linear nature of MLDM optimization our approach outperforms other volumetric segmentation methods in terms of speed.

Show publication details

Erdt, Marius; Kirschner, Matthias; Wesarg, Stefan

Smart Manual Landmarking of Organs

2010

Dawant, Benoit M. (Ed.) et al.: Medical Imaging 2010: Image Processing. Part One : Progress in Biomedical Optics and Imaging. Proceedings Issue. Vol. 11, No. 33. Bellingham: SPIE Press, 2010. (Proceedings of SPIE 7623), pp. 76234Y-1 - 76234Y-9

SPIE Medical Imaging Symposium <2010, San Diego, CA, USA>

Statistical shape models play a very important role in most modern medical segmentation frameworks. In this work we propose an extension to an existing approach for statistical shape model generation based on manual mesh deformation. Since the manual acquisition of ground truth segmentation data is a prerequisite for shape model creation, we developed a method that integrates a solution to the landmark correspondence problem in this particular step. This is done by coupling a user guided mesh adaptation for ground truth segmentation with a simultaneous real time optimization of the mesh in order to preserve point correspondences. First, a reference model with evenly distributed points is created that is taken as the basis of manual deformation. Afterwards the user adapts the model to the data set using a 3D Gaussian deformation of varying stiffness. The resulting meshes can be directly used for shape model construction. Furthermore, our approach allows the creation of shape models of arbitrary topology. We evaluate our method on CT data sets of the kidney and 4D MRI time series images of the cardiac left ventricle. A comparison with standard ICP-based and population-based optimization based correspondence algorithms showed better results both in terms of generalization capability and specificity for the model generated by our approach. The proposed method can therefore be used to considerably speed up and ease the process of shape model generation as well as remove potential error sources of landmark and correspondence optimization algorithms needed so far.

Show publication details

Rahman, Sami ur; Wesarg, Stefan

Upsampling of Cardiac MR Images: Comparison of Averaging and Super-Resolution for the Combination of Multiple Views

2010

University of Ioannina: ITAB 2010 : 10th International Conference on Information Technology and Applications in Biomedicine. New York: IEEE, Inc., 2010, 4 p.

International Conference on Information Technology and Applications in Biomedicine (ITAB) <10, 2010, Corfu, Greece>

In cardiac imaging quantitative analysis heavily depends on the quality of the available image data. Cardiac MRI provides highly anisotropic voxel data. By combining multiple orthogonal data sets, isotropic volume data can be reconstructed. In this paper we investigate the increase in image quality by the use of a super-resolution reconstruction (SRR) algorithm. In particular, we compare a simple averaging with an SRR for combining two and three orthogonal views, respectively. We show that SRR outperforms averaging in case of the combination of two views, but that in case of three views SRR has no additional benefit compared to averaging. We conclude that for cardiac images where motion of the organ plays an important role the prior alignment of the image data sets is a limiting factor for successfully applying an SRR.

Show publication details

Wesarg, Stefan; Kirschner, Matthias

3D Visualization of Medical Image Data Employing 2D Histograms

2009

Banissi, Ebad (Ed.) et al.: Proceedings Second International Conference in Visualisation. VIZ 2009 : Visualisation in Built and Rural Environments - BioMedical Visualisation - Geometric Modelling and Imaging. Los Alamitos, Calif.: IEEE Computer Society Conference Publishing Services (CPS), 2009, 6 p.

International Conference BioMedical Visualisation (MediViz09) <6, 2009, Barcelona, Spain>

Transfer functions (TF) are a means for improving the visualization of 3D medical image data. If in addition to intensity another property is employed, two-dimensional TFs can be specified. For this, 2D histograms are helpful. In this work we investigate how the property feature size can be used for the definition of 2D TFs and the visualization of medical image data. Furthermore, we compare this method to approaches that employ gradient magnitude as second property. From our experiments with several medical image data we conclude, that structure size enhanced 2D histograms are more intuitive. This is especially true in the clinical area, where physicians are much more familiar with the meaning of the size of anatomical structures than with the concept of gradient magnitude.

Show publication details

Kirschner, Matthias; Wesarg, Stefan

Area Preserving Parameterisation of Shapes with Spherical Topology

2009

Fischer, Stefan (Ed.) et al.: Informatik 2009. Im Focus das Leben : Beiträge zur 39. Jahrestagung der Gesellschaft für Informatik e.V.. Bonn: Gesellschaft für Informatik, 2009. (GI-Edition - Lecture Notes in Informatics (LNI) P-154), pp. 1257-1266

Informatik 2009 - Jahrestagung der Gesellschaft für Informatik e.V. (GI) <39, 2009, Lübeck, Germany>

Statistical shape models are powerful tools for model-based segmentation and have been successfully applied to the segmentation of various structures in medical images. Though the segmentation algorithms based on statistical shape models are simple, finding corresponding landmarks for the construction of the models is a challenging optimisation task. State-of-the-art algorithms that solve the correspondence problem require a representation of the training shapes in a suitable parameter space. The mapping of a shape to a parameter space can introduce large area distortions so that simple sampling techniques can not reconstruct the original shapes sufficiently well. In this paper, we propose an algorithm to construct area preserving parameterisations of shapes with spherical topology. Using our approach, good reconstructions of the original shapes can be achieved by uniform sampling. In contrast to previously published methods that use a black box optimisation approach, we exploit knowledge about the shortcomings of initial parameterisations.

Show publication details

Gutbell, Ralf; Wesarg, Stefan (Betreuer)

Fokussierte Visualisierung medizinischer Bilddaten

2009

Darmstadt, TU, Bachelor Thesis, 2009

In der medizinischen Visualisierung werden Ärzte oft mit großen Datensätzen konfrontiert aus denen sie ihre Diagnose stellen müssen. Die anatomische Struktur, welche untersucht werden soll, ist dabei meist von anderen anatomischen Elementen verdeckt und erschwert dem Arzt die Arbeit. Mehrere Arbeitsgruppen, wie zum Beispiel Luft et al. und Ebert et al., beschäftigen sich mittlerweile damit, die Volumen optisch, mit Hilfe von bekannten menschlichen Wahrnehmungsmustern, aufzubereiten, so dass dem Arzt bei der Diagnose geholfen wird. Die vorliegende Arbeit nutzt bekannte Aspekte der menschlichen Wahrnehmungen, wie den Fokus des Auges, um dem Arzt optische Hilfen zu geben. Dabei wird hauptsächlich, abhängig von einem Distanzmaß, die Opazität und Schärfe von Voxeln manipuliert. Die Bestimmung des Fokus geschieht durch den Benutzer, indem er den Fokus im Volumen markiert. Zusätzlich zum Fokus bestimmt der Benutzer noch einen zweiten Radius, welcher als äußerer Rand benutzt wird, um zu bestimmen wie weit außerhalb des Fokus das Volumen noch sichtbar sein soll. Zwischen dem Fokus und dem Radius werden zusätzlich zur manipulierten Opazität die Voxel auch noch verschmiert, um den Effekt zu verstärkten. Die Intensität des Blurrings kann wiederum vom Benutzer gesteuert werden. Die Evaluation mit mehreren verschiedenen Datensätzen ergab, dass der Ansatz ein mächtiges Werkzeug zur Untersuchung sein kann. Die Benutzbarkeit der momentanen Umsetzung ist noch stark von der Größe des Volumens abhängig, da sonst zu hohe Wartezeiten auftreten. Auch sehr homogene Volumen sind eine Schwäche der jetzigen Umsetzung, welche aber lösbare Probleme darstellen und deren Umsetzung Teil der zukünftigen Arbeiten sein wird.

Show publication details

Jung, Florian; Wesarg, Stefan (Betreuer)

GPU-basierte rigide Registrierung von 3D-Bilddaten

2009

Darmstadt, TU, Bachelor Thesis, 2009

Diese Bachelorarbeit beschäftigt sich mit der Frage, inwieweit sich der Prozess der rigiden Bildregistrierung auf der Grafikkarte implementieren lässt. Hierfür wird Nvidias CUDA verwendet. Zuerst wird der Algorithmus auf Parallelisierbarkeit untersucht und anschließend auf die Grafikkarte portiert, um so eine Beschleunigung gegenüber der CPU-Implementierung zu erreichen. Anschließend wird die Geschwindigkeit und Genauigkeit der beiden Implementierungen verglichen und auf unterschiedlichen Systemen getestet. Durch die Verwendung von CUDA konnte der Algorithmus um den Faktor 4-10 beschleunigt werden, abhängig von der verwendeten Hardware.

Show publication details

Wesarg, Stefan; Kirschner, Matthias

Gradient Magnitude vs. Feature Size: Comparing 2D Histograms for Transfer Function Specification

2009

Computer Graphics Society (CGS): CGI 2009 : CGI 2009. New York: ACM, 2009, pp. 115-119

Computer Graphics International (CGI) <2009, Victoria, BC, Canada>

Transfer function (TF) specification for 3D visualization can be improved by incorporating more than just intensity information. Gradient magnitude is an additional property that has been used predominantly for defining TFs based on 2D histograms. Recently, feature size has been introduced as an alternative. In this work we compare the usability of such 2D histograms for the visualization of 3D scalar data. By giving visualization examples for well-known volume data sets we conclude that employing feature size for TF specification is more intuitive than gradient magnitude based approaches.

Show publication details

Kirschner, Matthias; Wesarg, Stefan

Propagation of Shape Parameterisation for the Construction of a Statistical Shape Model of the Left Ventricle

2009

Magnor, Marcus A. (Ed.) et al.: Proceedings of the Vision, Modeling, and Visualization Workshop 2009 : VMV 2009. Magdeburg: Otto-von-Guericke-Universität, 2009, p. 313-320

Vision, Modeling, and Visualization Workshop (VMV) <14, 2009, Braunschweig, Germany>

Statistical Shape Models (SSMs) have been successfully applied to both segmentation and the description of the dynamic behaviour of the heart. SSMs are learned from a set of training examples, which are represented by vectors of corresponding landmarks. While the construction of a SSMis simple when a landmark representation of the training shapes is available, the extraction of corresponding landmarks from training images or meshes of different sizes is difficult. Optimisation schemes that solve this so-called correspondence problem rely on a parameter space representation of the input shapes. These optimisation schemes tend to be sensitive to the initial parameterisation of the input shapes. In this work, we present an algorithm to produce a consistent spherical parameterisation for shapes of the left ventricle. Our algorithm propagates the spherical parameterisation of a root shape within seconds to all other shapes. We demonstrate the effectiveness of our approach by extracting a SSM from the parameterisations generated by our algorithm.

Show publication details

Erdt, Marius; Kirschner, Matthias; Wesarg, Stefan

Simultaneous Segmentation and Correspondence Establishment for Statistical Shape Models

2009

Magnenat-Thalmann, Nadia (Ed.): Modelling the Physiological Human : 3D Physiological Human Workshop. Proceedings. Berlin, Heidelberg, New York: Springer, 2009. (Lecture Notes in Computer Science (LNCS) 5903), pp. 25-35

3D Physiologigal Human Workshop (3DPH) <2, 2009, Zermatt, Switzerland>

Statistical Shape Models have been proven to be valuable tools for segmenting anatomical structures of arbitrary topology. Being based on the statistical description of representative shapes, an initial segmentation is required - preferably done by an expert. For this purpose, mostly manual segmentation methods followed by a mesh generation step are employed. A prerequisite for generating the training data based on these segmentations is the establishment of correspondences between all training meshes. While existing approaches decouple the expert segmentation from the correspondence establishment step, we propose in this work a segmentation approach that simultaneously establishes the landmark correspondences needed for the subsequent generation of shape models. Our approach uses a reference segmentation given as a regular mesh. After an initial placement of this reference mesh, it is manually deformed in order to best match the boundaries of the considered anatomical structure. This deformation is coupled with a real time optimization that preserves point correspondences and thus ensures that a pair of landmark points in two different data sets represents the same anatomical feature. We applied our new method to different anatomical structures: vertebra of the spinal chord, kidney, and cardiac left ventricle. In order to perform a visual evaluation of the degree of correspondence between different data sets, we have developed well adapted visualization methods. From our tests we conclude that the expected correspondences are established during the manual mesh deformation. Furthermore, our approach considerably speeds up the shape model generation, since there is no need for an independent correspondence establishment step. Finally, it allows the creation of shape models of arbitrary topology and removes potential error sources of landmark and correspondence optimization algorithms needed so far.

Show publication details

Wesarg, Stefan; Kirschner, Matthias

Structure Size Enhanced Histogram: A Transfer Function for 3D Volume Visualization

2009

Meinzer, Hans-Peter (Ed.) et al.: Bildverarbeitung für die Medizin 2009. Proceedings : Algorithmen, Systeme, Anwendungen. Berlin; Heidelberg; New York: Springer, 2009. (Informatik aktuell), pp. 16-20

Workshop Bildverarbeitung für die Medizin (BVM) <12, 2009, Heidelberg, Germany>

Direct volume visualization requires the definition of transfer functions (TFs) for the assignment of opacity and color. Multi-dimentional TFs are based on at least two image properties, and are specified by means of 2D histograms. In this work we propose a new type of a 2D histogram which combines grey value with information about the size of the structures. This Structure Size Enhanced (SSE) histogram is an intuitive approach for representing anatomical features. Clinicians - the users we are focusing on - are much more familiar with selecting features by their size than by their gradient magnitude value. As a proof of concept, we employ the SSE histogram for the definition of two-dimentional TFs for the visualization of 3D MRI and CT image data.

Show publication details

Wesarg, Stefan; Khan, M. Fawad; Jähne, Marion; Lacalli, Christina

Automatisierte Analyse der Koronararterien basierend auf MSCT-Daten

2008

Deutsche Zeitschrift für klinische Forschung: DZKF, Vol.11 (2008), 3/4, pp. 28-33

Ein frühzeitiges Erkennen von Gefäßverengungen ist von großer Wichtigkeit für die Diagnose Koronarer Herzkrankheiten. Durch automatisierte Koronaranalysen können zuverlässige Ergebnisse nun bei einem deutlich reduzierten Zeitaufwand geliefert werden.

Show publication details

Lacalli, Christina; Jähne, Marion; Wesarg, Stefan

Automatisierte Verfahren zur verbesserten Visualisierung der Koronararterien in MSCT-Daten und für die direkte Vergleichbarkeit zur Angiographie

2008

Tolxdorff, Thomas (Ed.) et al.: Bildverarbeitung für die Medizin 2008. Proceedings : Algorithmen, Systeme, Anwendungen. Heidelberg: Springer, 2008. (Informatik aktuell), pp. 283-287

Workshop Bildverarbeitung für die Medizin (BVM) <11, 2008, Berlin, Germany>

In diesem Beitrag stellen wir neue, automatisierte Verfahren zur Visualisierung der Koronararterien einerseits und für eine direkte Vergleichbarkeit mit konventionellen Angiogrammen andererseits vor. Unser Ansatz umfasst Methoden für die automatische Extraktion des Herzens aus kontrastverstärkten CT-Daten, sowie für die Maskierung grosser kontrastmittelgefüllter Kavitäten des Herzens, um die Sichtbarkeit der Koronararterien bei der Darstellung mittels Volumenrendering zu verbessern. Für den direkten Vergleich zur konventionellen Angiographie haben wir ein Verfahren zur automatischen Generierung von - den Angiogrammen entsprechenden - Projektionsansichten aus den CTDaten entwickelt.

Show publication details

Wesarg, Stefan; Lacalli, Christina

Computation and Visualization of Asynchronous Behavior of the Heart

2008

Botha, C. P. (Ed.) et al.: Eurographics Workshop on Visual Computing for Biomedicine : EG VCBM 2008. Aire-la-Ville: Eurographics Association, 2008, pp. 29-36

Eurographics Workshop on Visual Computing for Biomedicine (VCBM) <1, 2008, Delft, Netherlands>

Nowadays, computer-aided diagnosis is widely used in the analysis of cardiac image data. Especially, for the investigation of the dynamic behavior of the heart, automated analysis tools for 4D data sets have been developed. A small set of descriptors of the heart's dynamics are established and used in the clinical routine. However, there exists a whole lot more of such parameters that can be extracted by analyzing 4D data sets. But, many of them are not used due to several reasons: time-consuming computation, no intuitive meaning, little clinical relevance, etc.. In this work we propose a novel descriptor for the dynamic behavior of the heart that can easily be computed from 4D data sets. It describes to which extent the heart exhibits an asynchronous movement. This novel descriptor ASYNCHRONISM is based on the already established measures WALL MOTION and WALL THICKENING, but reveals new, valuable information that is not available when relying only upon the two aforementioned parameters. The ASYNCHRONISM has an intuitive meaning, since it corresponds to the clinical classification scheme of wall motion abnormalities. Beyond its computation we present in this work also methods for its visualization as well as first preliminary results for 4D cardiac magnetic resonance image data.

Show publication details

Jähne, Marion; Lacalli, Christina; Wesarg, Stefan

Novel Techniques for Automatically Enhanced Visualization of Coronary Arteries in MSCT Data and for Drawing Direct Comparisons to Conventional Angiography

2008

Institute for Systems and Technologies of Information, Control and Communication (INSTICC): VISIGRAPP 2008. Proceedings : International Joint Conference on Computer Vision and Computer Graphics Theory and Applications [CD-ROM]. INSTICC Press, 2008, pp. 290-296

International Conference on Computer Vision Theory and Applications (VISAPP) <3, 2008, Funchal, Madeira, Portugal>

The new generation of multi-slice computed tomography (MSCT) scanners enables the radiologist to assess the coronary arteries in a non-invasive way. The question of particular interest is whether the quality of the findings based on MSCT data can compete with the gold standard - the coronary angiography. In this work we present novel automated methods for a reliable visualization of coronary arteries and for drawing direct visual side-by-side comparisons to conventional angiograms. Our approach comprises a new method for automatically extracting the heart from cardiac CT data and an advanced masking method for eliminating large cardiac cavities to obtain a better visibility of the coronary arteries in the rendered CT data. For drawing direct side-by-side comparisons we present a novel approach for simulating the conventional coronary angiography in an easy-to-handle manner. The new methods have been developed for and tested with contrast-enhanced cardiac CT datasets.

Show publication details

Wesarg, Stefan; Fellner, Dieter W. (Betreuer); Giannitsis, Evangelos (Betreuer)

Automatisierte Analyse und Visualisierung der Koronararterien und großen Kavitäten des Herzens für die klinische Anwendung

2007

Darmstadt, TU, Diss., 2007

Es werden verschiedene Verfahren zur Analyse von Bilddaten des kardiovaskulären Systems behandelt. Damit wird eine Verbesserung sowohl der Diagnose als auch der Planung von eventuell notwendigen Eingriffen erreicht. Die beschriebenen Verfahren zeichnen sich durch eine hohe Automatisierung und Reproduzierbarkeit der Analyseergebnisse sowie ihre klinische Anwendbarkeit aus. Augenmerk wird vor allem auf die auf der Oberfläche des Herzens liegenden Herzkranzgefäße und die den Hauptteil des Herzens bildenden Kavitäten - das linke und rechte Ventrikel gelegt. Hier werden verschiedene im Rahmen dieser Arbeit entwickelte, neue Segmentierungs- und Analyseverfahren vorgestellt und diskutiert. Im Falle der Herzkranzgefäße sind das trackingbasierte Segmentierungsansätze, die die Basis bilden für eine Analyse des Gefäßes hinsichtlich - der Detektion und Vermessung von Stenosen - des Vorhandenseins von harten Arterienverkalkungen - der Zusammensetzung des umliegenden Gewebes Desweiteren wird ein Verfahren vorgestellt, das es ermöglicht, die damit erreichten Ergebnisse mit der Koronarangiographie - dem Gold-Standard - zu vergleichen. Für eine angepaßte Präsentation der Analyseergebnisse werden speziell entwickelte Verfahren für deren optimale Visualisierung als auch die der Bilddaten selbst vorgestellt. Letztere betreffend wird ein automatisches Verfahren eingeführt, mit dessen Hilfe sich Strukturen wie der Brustkorb ausmaskieren lassen, die die direkte Sicht auf das Herz stören. Für die Analyse von linkem (LV) und rechtem Ventrikel (RV) werden automatisierte Segmentierungsverfahren vorgestellt, aus deren Ergebnis sich die die Dynamik der Ventrikel beschreibenden physikalischen Parameter ableiten lassen. Für das LV wird eine umfassende, automatische und detaillierte Analyse der Wandbewegung, Wanddickenzunahme und Volumenänderung vorgestellt. Als neuer Deskriptor für die Dynamik wird die Asynchronität eingeführt. Die für das LV entwickelten Analyseverfahren werden auf das RV übertragen und ermöglichen so eine ganz neue Qualität dessen Analyse. Die Präsentation der berechneten Parameter erfolgt in einer standardisierten Weise entsprechend den Empfehlungen der American Heart Association. Als Erweiterung dieser Darstellungsmöglichkeit wird die direkte Visualisierung dieser Größen zusammen mit einem 3D-Rendering des LV eingeführt. Dies fließt ein in eine kombinierte Darstellung von dynamischen Parametern und Infarktbereichen des Herzens. Letztere werden zudem automatisch quantifiziert. Die wesentlichen Fortschritte dieser Arbeit sind: 1. die Entwicklung zweier neuer trackingbasierter Algorithmen für die Segmentierung von Koronararterien in kontratverstärkten CT-Daten, 2. die Einführung neuer Visualisierungsmethoden für die Präsentation der Ergebnisse der Koronaranalyse, 3. die Schaffung direkter Vergleichsmöglichkeiten zwischen CT-Angiographie und konventioneller Angiographie, 4. die Kombination bestehender Segmentierungstechniken mit anatomischer Kenntnis für eine automatisierte Extraktion von linkem und rechtem Ventrikel, 5. die Etablierung von umfassenden Analyseverfahren für die Dynamik des linken Ventrikels, 6. die erstmalige Anwendung dieser Ansätze auf die Dynamik des rechten Ventrikels, 7. die Einführung eines das asynchrone Verhalten von Bereichen des Ventrikels beschreibenden neuen Parameters in die LV-Analyse und 8. die Erweiterung der Diagnostik von Infarktbereichen um eine automatische Narbenquantifizierung und neue Visualisierungsmethoden.

Show publication details

Wald, Diana; Wesarg, Stefan; Nowak, Stefanie

Quantifizierung und Visualisierung von Narbenbereichen des Myokards

2007

Horsch, Alexander (Ed.) et al.: Bildverarbeitung für die Medizin 2007. Proceedings : Algorithmen, Systeme, Anwendungen. Heidelberg: Springer, 2007. (Informatik aktuell), pp. 323-327

Workshop Bildverarbeitung für die Medizin (BVM) <10, 2007, München, Germany>

In diesem Beitrag wird ein automatisches Verfahren zur Quantifizierung von Narbenbereichen des Myokards aus MRT-Daten vorgestellt. Spezielles Augenmerk wird dabei auf die automatische Berechnung und die praxisnahe Präsentation der Analysedaten gelegt. Weiterhin werden dem Anwender verschiedene Möglichkeiten geboten, die Ergebnisse auf ihre Korrektheit zu überprüfen und wenn nötig zu korrigieren, um Fehldiagnosen zu vermeiden.

Show publication details

Basdogan, Cagatay; Sedef, Mert; Harders, Matthias; Wesarg, Stefan

VR-Based Simulators for Training in Minimally Invasive Surgery

2007

IEEE Computer Graphics and Applications, Vol.27 (2007), 2, pp. 54-66

Simulation-based training using VR techniques is a promising alternative to traditional training in minimally invasive surgery (MIS). Simulators let the trainee touch, feel, and manipulate virtual tissues and organs through the same surgical tool handles used in actual MIS while viewing images of tool-tissue interactions on a monitor as in real laparoscopic procedures.

Show publication details

Nowak, Stefanie; Wesarg, Stefan (Betreuer)

4D Narbendarstellung des Herzens Visualisierung und Modellierung der Dynamik und Funktion des linken Ventrikels

2006

Siegen, Univ., Diplomarbeit, 2006

Über medizinische Bildgebungsverfahren ist es Ärzten möglich, einen Einblick in den menschlichen Körper und seine komplexen Vorgänge zu erhalten und so Diagnose und Behandlungsplanung zu erstellen. Besonders nach einem Herzinfarkt ist eine Abschätzung des Ausmaßes der Beeinträchtigung der Herzfunktion notwendig. Dafür, aber auch für eine prophylaktische Einschätzung der Herztätigkeit, können Parameter berechnet werden, die die Funktion des Herzens beschreiben. Aufgrund der Masse an Bilddaten, die für eine solche Berechnung anfallen, sind automatische, intuitiv bedienbare Applikationen notwendig, um den Arzt bei seiner Diagnose zu unterstützen. Die vorliegende Arbeit beschäftigt sich mit der drei- und vierdimensionalen Darstellung der Funktion und Deformation des linken Ventrikels bei Herzinfarktpatienten. Eine Visualisierung der Ventrikelfunktion mit Überlagerung des vernarbten Gewebes illustriert die durch den Infarkt hervorgerufenen Bewegungsstörungen. Untersucht wurde der Zusammenhang der Lokalisierung des vernarbten Gewebes zu Bereichen mit kritischen Funktionsparametern. Dazu wurde ein automatisches Verfahren entwickelt, dass ein dreidimensionales, die Funktionsparameter farbkodierendes Modell des linken Ventrikels zu jeder Herzphase erstellt und zudem einen Satz von Bewegungsvektoren über den Herzzyklus ableitet. Die Deformationsfelder werden für eine künstliche Deformation auf den nur zu einem Zeitpunkt im Herzzyklus vorhandenen statischen Datensatz (late enhancement) angewendet. Es zeigt sich, dass neben einer intuitiven Erkenntnis der Bewegungsstörungen durch den in einer Animation simulierten Pumpvorgang, die überlagerte Visualisierung der Deformation des Narbendatensatzes dem Arzt eine kompakte Informationsdarstellung bietet. Über direkte Visualisierungstechniken wird der Narbenbereich zufrieden stellend hervorgehoben visualisiert. Für eine Darstellung und Beurteilung der Qualtät des entwickelten Verfahrens, wurde es an Daten von 12 Herzinfarktpatienten getestet und ausgewertet.

Show publication details

Wesarg, Stefan; Nowak, Stefanie

An Automated 4D Approach for Left Ventricular Assessment in Clinical Cine MR Images

2006

Hochberger, Christian (Ed.) et al.: Informatik 2006. Informatik für Menschen. Band 1 : Beiträge zur 36. Jahrestagung der Gesellschaft für Informatik e.V.. Bonn: Gesellschaft für Informatik, 2006. (GI-Edition - Lecture Notes in Informatics (LNI) P-93), pp. 483-490

Informatik 2006 - Jahrestagung der Gesellschaft für Informatik e.V. (GI) <36, 2006, Dresden, Germany>

CineMagnetic Resonance (MR) imaging has become the method-of-choice for the examination of the dynamic behaviour of the heart. An assessment of the left ventricle can reveal regions of myocardial dysfunction and their severeness. The scope of this work is a complete analysis of the left ventricular dynamics for the usage in a clinical environment. For that purpose, endocardial and epicardial borders are automatically extracted in 3D cine data in a first step. This is followed by a segmentation of the endocardium and the myocardium into 17 segments following the recommendations of the American Heart Association and the computation of common global volumetric values (stroke volume, ejection fraction etc.) and parameters that describe the left ventricular dynamics (wall motion, wall thickening). A retrospective analysis of cardiac cine MR image data from 20 patients (healthy ones, patients with abnormal wall motion, and patients who suffered an infarction) has been done. That image data has been acquired in the clinical routine at two different hospitals. The here presented automated approach led to a successful segmentation and assessment of the left ventricle for all data sets. The pathological cases could be identified easily due to their characteristic change of the motion pattern. The main advantage of our approach is the reproducibility of the assessment results and the gain in time for the cardiologist who has to analyse the huge amount of cine data.

Show publication details

Khan, M. Fawad; Wesarg, Stefan; Gurung, Jessen; Dogan, Selami; Maataoui, Adel; Brehmer, Boris; Herzog, Christopher; Ackermann, Hanns; Aßmus, Birgit; Vogl, Thomas J.

Facilitating Coronary Artery Evaluation in MDCT Using a 3D Automatic Vessel Segmentation Tool

2006

European Radiology, Vol.16 (2006), 8, pp. 1789 - 1795

The purpose of this study was to investigate a 3D coronary artery segmentation algorithm using 16-row MDCT data sets. Fifty patients underwent cardiac CT (Sensation 16, Siemens) and coronary angiography. Automatic and manual detection of coronary artery stenosis was performed. A 3D coronary artery segmentation algorithm (Fraunhofer Institute for Computer Graphics, Darmstadt) was used for automatic evaluation. All significant stenoses (>50%) in vessels >1.5 mm in diameter were protocoled. Each detection tool was used by one reader who was blinded to the results of the other detection method and the results of coronary angiography. Sensitivity and specificity were determined for automatic and manual detection as well as was the time for both CT-based evaluation methods. The overall sensitivity and specificity of the automatic and manual approach were 93.1 vs. 95.83% and 86.1 vs. 81.9%. The time required for automatic evaluation was significantly shorter than with the manual approach, i.e., 246.04±43.17 s for the automatic approach and 526.88±45.71 s for the manual approach (P<0.0001). In 94% of the coronary artery branches, automatic detection required less time than the manual approach. Automatic coronary vessel evaluation is feasible. It reduces the time required for cardiac CT evaluation with similar sensitivity and specificity as well as facilitates the evaluation of MDCT coronary angiography in a standardized fashion.

Show publication details

Wesarg, Stefan; Khan, M. Fawad; Firle, Evelyn

Localizing Calcifications in Cardiac CT Data Sets Using a New Vessel Segmentation Approach

2006

Journal of Digital Imaging, Vol.19 (2006), 3, pp. 249-257

The new generation of multislice computed tomography (CT) scanners allows for the acquisition of high-resolution images of the heart. Based on that image data, the heart can be analyzed in a noninvasive way-improving the diagnosis of cardiovascular malfunctions on one hand, and the planning of an eventually necessary intervention on the other. One important parameter for the evaluation of the severeness of a coronary artery disease is the number and localization of calcifications (hard plaques). This work presents a method for localizing these calcifications by employing a newly developed vessel segmentation approach. This extraction technique has been developed for, and tested with, contrast-enhanced CT data sets of the heart. The algorithm provides enough information to compute the vessel diameter along the extracted segment. An approach for automatically detecting calcified regions that combines diameter information and gray value analysis is presented. In addition, specially adapted methods for the visualization of these analysis results are described.

Show publication details

Lacalli, Christina; Wesarg, Stefan (Betreuer)

Matching von Angiographiebildern mit tomographischen Volumendaten

2006

Koblenz/Landau, Univ., Diplomarbeit, 2006

Die Koronararterien sind verantwortlich für die Versorgung des Herzmuskels mit Sauerstoff und Nährstoffen. Verengungen oder Verstopfungen dieser Gefäße können Herz-Kreislauf-Erkrankungen bis hin zum Herzinfarkt auslösen. Gegenwärtig stellt die invasive Koronarangiographie den Goldstandard innerhalb der Diagnostik dieser Erkrankungen dar. Eine alternative, nicht-invasive Untersuchungsmöglichkeit ist durch die Mehrschicht-Spiral-Computertomographie (MSCT) gegeben. Seit Einführung dieses Verfahrens 1998 ist die Fragestellung nach der Korrelation der aus den Volumendaten gewonnen Informationen über die Koronararterien mit denen aus der Koronarangiographie aufgekommen. In dieser Arbeit wurde ein Matchingverfahren entwickelt, das es dem Mediziner erstmals ermöglicht, mit geringem Zeitaufwand und minimaler Interaktion, einzelne Angiogramme mit entsprechenden CT-Datensätzen des Thorax in Deckung zu bringen, um so Informationen über das Koronararteriensystem aus beiden bildgebenden Verfahren direkt miteinander vergleichen zu können. Das Verfahren basiert auf der Generierung künstlicher Projektionsbilder aus CT-Datensätzen des Thorax, sogenannter Digitally Reconstructed Radiographies (DRRs), und einer anschließenden Registrierung dieser mit den entsprechenden Koronarangiogrammen. Das entwickelte Verfahren umfasst darüberhinaus einen neuen Ansatz, um kontrastverstärkte Herzstrukturen, die die Sicht auf die Koronararterien in den Projektionsbildern stören, auszumaskieren. Dieser vorverarbeitende Schritt bildet eine wesentliche Grundlage, um eine anschließende Registrierung überhaupt durchführen zu können. Die Visualisierung der Ergebnisse erfolgt durch eine fusionierte Darstellung der beiden Bildmodalitäten, sowie durch eine Parallelansicht der registrierten Datensätze.

Show publication details

Khan, M. Fawad; Dogan, Selami; Maataoui, Adel; Wesarg, Stefan; Gurung, Jessen; Ackermann, Hanns; Schiemann, Mirko; Wimmer-Greinecker, Gerhard; Vogl, Thomas J.

Navigation-Based Needle Puncture of a Cadaver Using a Hybrid Tracking Navigational System

2006

Investigative Radiology, Vol.41 (2006), 10, pp. 713-720

Purpose: The purpose of this study was to determine the puncture accuracy of a navigational system, Medarpa, in a soft tissue environment using augmented overlay imaging. Materials and Methods: Medarpa is an optical electromagnetic tracking system, which allows tracking of instruments, the radiologist's head position, and the transparent display. The display superimposes a computed tomography scan of a cadaver chest on a human cadaver in real time. In group A, needle puncture was performed using the Medarpa system. Three targets located inside the cadaver chest were selected. In group B, the same targets were used to perform standard computed tomography-guided puncture using a single-slice technique. A total of 42 punctures were performed in each group. Postpuncture computed tomography scans were made to verify needle tip positions. Results: Mean deviation from targets was 8.42 mm ± 1.78 mm for group A and 8.90 mm ± 1.71 mm for group B. No significant difference was found between group A and B in any target (P > 0.05). No significant difference was found between the targets of the same group (P > 0.05). Procedural time for 42 punctures was 160 minutes in group A versus 289 minutes in group B (P < 0.05). Conclusion: Needle puncture in a soft tissue environment using the navigational system Medarpa can be reliably performed and matches the accuracy achieved by a computed tomography-guided puncture technique.

Show publication details

Wald, Diana; Wesarg, Stefan (Betreuer)

Quantifizierung und Visualisierung von Narbenbereichen des Myokards

2006

Koblenz/Landau, Univ., Diplomarbeit, 2006

Die vorliegende Diplomarbeit thematisiert die quantitative Analyse und die Visualisierung von Infarktgewebe des linken Herzmuskels. Im Mittelpunkt der Untersuchung steht das Ausmaß der Narbe und deren Deformation über den Herzzyklus. Für die Narbenausdehnung stehen so genannte Late-Enhancement-Daten zur Verfügung, die das avitale Gewebe durch ein Kontrastmittel hervorheben. Anhand von automatisierten Verfahren wird die Narbe aus den Bilddaten extrahiert und auf ihre Größe, Lokalisation und Transmuralität quantifiziert. Die Transmuralität gibt dabei das lokale Verhältnis zwischen der Herzwand- und der Narbenbreite an. Des Weiteren wird die Narbe für die Beurteilung der Beschaffenheit dreidimensional in dem Analysefenster dargestellt. Der Mediziner kann durch das entwickelte Verfahren innerhalb kürzester Zeit Aussagen über das Ausmaß und den Ursprung des Herzinfarktes treffen und zudem die Ergebnisse durch verschiedene visuelle Darstellungen kontrollieren. Die Deformation des Narbengewebes über den Herzzyklus und deren Integration mit den dynamischen Cine-Daten wurde bereits in einer vorangegangenen Diplomarbeit umgesetzt. Im Rahmen dieser Arbeit wird eine visuelle Verbesserung der Deformationsergebnisse angestrebt, die das Narbengewebe aus den Volumendaten extrahiert. Das avitale Gewebe wird durch das Eliminieren von uninteressanten Bildinformationen hervorgehoben und verbessert somit die visuelle Analyse der Narbendeformation über den Herzzyklus. Beide Verfahren liefern eine detaillierte und eindeutige Analyse des Infarktgewebes, die die manuelle Untersuchung in der klinischen Praxis ergänzen kann.

Show publication details

Dold, Christian; Wesarg, Stefan; Firle, Evelyn; Seitel, Mathias

4D-Segmentierung des linken Ventrikels basierend auf Region Growing und einer speziellen Bildaufbereitung angewendet auf CT, MR und U/S

2005

Meinzer, Hans-Peter (Ed.) et al.: Bildverarbeitung für die Medizin 2005. Proceedings : Algorithmen, Systeme, Anwendungen. Heidelberg: Springer, 2005. (Informatik aktuell), pp. 133-137

Workshop Bildverarbeitung für die Medizin (BVM) <2005, Heidelberg, Germany>

Es wurde eine Softwareplattform entwickelt, die 4D-Datensätze von CT-, MR- und Ultraschalldaten segmentieren kann. Dabei ist sie sehr leicht zu bedienen und auf die unterschiedlichen Modalitäten anpassbar. Bei der Herzanalyse wird enorm viel Zeit gespart, da lediglich die Region-of-Interest, ein Seedpunkt und die Ober-/ Untergrenze des Grauwertes angegeben werden muß, um eine 4D-Segmentierung vorzunehmen. Die Daten müssen nicht, wie üblich, in der Kurzachsenaufnahme zur Verfügung stehen, da sie beliebig vom Radiologen mit der Software ausgerichtet werden können. Auch ist das Ergebnis nicht mehr von der zu bedienenden Person abhängig sondern reproduzierbar.

Show publication details

Khan, M. Fawad; Dogan, Selami; Maataoui, Adel; Gurung, Jessen; Schiemann, Mirko; Ackermann, Hanns; Wesarg, Stefan; Sakas, Georgios; Vogl, Thomas J.

Accuracy of Biopsy Needle Navigation Using the Medarpa System-computed Tomography Reality Superimposed on the Site of Intervention

2005

European Radiology, Vol.15 (2005), 11, pp. 2366-2374

The aim of this work was to determine the accuracy of a new navigational system, Medarpa, with a transparent display superimposing computed tomography (CT) reality on the site of intervention. Medarpa uses an optical and an electromagnetic tracking system which allows tracking of instruments, the radiologist and the transparent display. The display superimposes a CT view of a phantom chest on a phantom chest model, in real time. In group A, needle positioning was performed using the Medarpa system. Three targets (diameter 1.5 mm) located inside the phantom were punctured. In group B, the same targets were used to perform standard CT-guided puncturing using the single-slice technique. The same needles were used in both groups (15 G, 15 cm). A total of 42 punctures were performed in each group. Post puncture, CT scans were made to verify needle tip positions. The mean deviation from the needle tip to the targets was 6.65±1.61 mm for group A (range 3.54-9.51 mm) and 7.05± 1.33 mm for group B (range 4.10-9.45 mm). No significant difference was found between group A and group B for any target (p>0.05). No significant difference was found between the targets of the same group (p>0.05). The accuracy in needle puncturing using the augmented reality system, Medarpa, matches the accuracy achieved by CT-guided puncturing technique.

Show publication details

Wesarg, Stefan

AHA Conform Analysis of Myocardial Function Using and Extending the Toolkits ITK and VTK

2005

Lemke, Heinz U. (Ed.) et al.: Computer Assisted Radiology and Surgery. Proceedings : CARS 2005. Amsterdam: Elsevier, 2005. (International Congress Series 1281), pp. 44-49

Computer Assisted Radiology and Surgery (CARS) <19, 2005, Berlin, Germany>

Modern dynamic image acquisition techniques allow for a non-invasive imaging of the beating heart. Thus, ischemic areas of the heart's wall can be detected by anlyzing the left ventricular myocardial function. For a standardization of that task the American Heart Association (AHA) has published recommendations for the myocardial segmentation and nomenclature. We propose in this paper a comprehensive, fully automatic, and AHA conform analysis of the myocardial function that is based on the segmentation results of the left ventricular cavity for the whole cardiac cycle. The physical values wall motion, ejection fraction, and wall thickening are computed and visualized employing and extending the toolkits ITK and VTK. First test results using dynamic MRI data are presented.

Show publication details

Wesarg, Stefan; Dold, Christian; Tadge, Tonio; Seitel, Mathias

Analyse des linken Ventrikels nach AHA-Richtlinien unter Verwendung von VTK

2005

Meinzer, Hans-Peter (Ed.) et al.: Bildverarbeitung für die Medizin 2005. Proceedings : Algorithmen, Systeme, Anwendungen. Heidelberg: Springer, 2005. (Informatik aktuell), pp. 16-20

Workshop Bildverarbeitung für die Medizin (BVM) <2005, Heidelberg, Germany>

Die Analyse des linken Ventrikels basierend auf dynamischen 4D-Bilddaten (3D-Volumen + Zeit) stellt für die Kardiologie eine wichtige Methode zur Diagnose von Herzproblemen dar. Durch eine Untersuchung der Wandbewegung und der damit verbundenen Volumenänderung lassen sich ischämische Bereiche des Herzens erkennen. Zur einheitlichen Darstellung der Ergebnisse hat die American Heart Association (AHA) die Verwendung eines Bull's-Eye-Displays empfohlen. Diese Arbeit präsentiert eine vollautomatische Analyse des linken Ventrikels basierend auf dessen Segmentierung. Dabei werden die die Wandbewegung und die Volumenänderung beschreibenden Größen aus den Ergebnissen einer vorangehenden Segmentierung des Ventrikels berechnet und AHA-konform dargestellt. Zur Visualisierung wird dabei das Toolkit VTK verwendet, das um ein Bull's-Eye-Display-Widget erweitert wird.

Show publication details

Jähne, Marion; Wesarg, Stefan (Betreuer)

Automatische Extraktion des Herzens aus kontrastverstärkten Schichtaufnahmen des Thorax

2005

Darmstadt, TU, Diplomarbeit, 2005

In der Herzchirurgie bietet der Einsatz computerunterstützer Verfahren erheblich verbesserte Möglichkeiten zur Planung und Durchführung einer Operation. Die Segmentierung des Herzens bildet dabei einen Vorverarbeitungsschritt für die nachfolgende Visualisierung oder problemorientierte Segmentierung. Durch die Segmentierung des Herzens aus den Thorax-Aufnahmen werden Strukturen, die die Sicht auf das Herz stören, unterdrückt, so dass man eine uneingeschränkte Sicht auf die Herzstrukturen erhält. In dem hier entworfenen Verfahren erfolgt die Segmentierung des Herzens durch Kombination aus einem Schwellwertverfahren (Otsu-Verfahren), mit dem der Datensatz vorsegmentiert wird, und einem 2D-Suchstrahl-Verfahren zur Erkennung der Herzoberfläche in den axialen Schichten. Die Suchstrahlen finden die klaren hell-dunkel Übergänge vom Herzen zum umliegenden Gewebe. An den Stellen, an denen das Herz an das Brustbein oder die Aorta grenzt, werden die zu langen Suchstrahlen interpoliert. Das Verfahren wurde auf 25 CT- und 2 MR-Datensätzen ausgewertet. Die durchschnittliche Bearbeitungszeit beträgt auf einem dualen AMD Athlon mit 2 GHz je nach Größe des Datensatzes zwischen 5 und 9 Sekunden.

Show publication details

Wesarg, Stefan; Firle, Evelyn

CT-basierte Analyse von Koronararterien zur Unterstützung eines TECAB-Grafting

2005

Meinzer, Hans-Peter (Ed.) et al.: Bildverarbeitung für die Medizin 2005. Proceedings : Algorithmen, Systeme, Anwendungen. Heidelberg: Springer, 2005. (Informatik aktuell), pp. 26-30

Workshop Bildverarbeitung für die Medizin (BVM) <2005, Heidelberg, Germany>

Verkalkungen der Herzkranzgefäße können zu Erkrankungen des kardio-vaskulären Systems bis hin zu einem Herzinfarkt führen. Eine gängige Behandlungsmethode ist die Bypaß-Operation, welche mit Hilfe moderner Telemanipulator-Systeme minimal-invasiv durchgeführt werden kann. Dies ist allerdings mit Einschränkungen für den Chirurgen bezüglich der Sicht auf die Koronararterien und des Zugang zu selbigen verbunden. Diese Arbeit beschreibt neuartige Methoden für die Unterstützung eines TECAB-Grafting, einer speziellen Form einer minimal-invasiven Bypaß-Operation. Basierend auf einer Segmentierung des Herzkranzgefäßes werden harte Arterienverkalkungen automatisch detektiert und die Umgebung des Gefäßes analysiert. Zudem werden speziell entwickelte Methoden für die Visualisierung der Analyseergebnisse vorgestellt.

Show publication details

Keil, Matthias; Firle, Evelyn (Betreuer); Wesarg, Stefan (Betreuer)

Entwurf eines multimodalen Volumenvisualisierungssystems

2005

Magdeburg, Univ., Diplomarbeit, 2005

Durch medizinische Bildgebung ist es den Ärzten und Forschern möglich geworden, Einblick ins Innere des menschlichen Körpers zu bekommen. Eine gemeinsame Darstellung von registrierten Volumendatensätzen des gleichen Objektes, welche zudem mit verschiedenen Modalitäten aufgenommen wurden (multimodale Visualisierung), ermöglicht die Nutzung der Vorteile der einzelnen bildgebenden Verfahren und ebenso eine Kompensation der Nachteile. Weiterhin können komplementäre Informationen, die sich aus den speziellen Abbildungseigenschaften der Verfahren ergeben, gemeinsam präsentiert werden. Eine dreidimensionale Visualisierung bietet zudem einen natürlichen Blick auf das Objekt. Für diesen Zweck wurde ein multimodales Volumenvisualisierungssystem für die dreidimensionale Darstellung registrierter medizinischer Datensätze entworfen und implementiert. Für die Darstellung der Bilddaten wurde eine direkte Volumenvisualisierung mittels mehrdimensionalen Transferfunktionen (TFs) für Farbe und Transparenz gewählt. Die Grundlage für mehrdimensionale TFs bilden Histogramme über die Intensität und den Gradientenbetrag der Bilddaten. Durch die Verwendung von Spatialized Transfer Functions und die damit verbundene halbautomatische Klassifikation des 2D-Histogramms in Bereiche, welche Strukturen im Volumendatensatz repräsentieren, gelang es den Spezifikationsaufwand für den Nutzer auf die Auswahl dieser Bereiche zu reduzieren. Das Konzept des Visualisierungssystems umfasst zwei Ansätze. Im sogenannten multiintensitätsbasierten Ansatz definiert der Nutzer eine gemeinsame TF für die beiden darzustellenden Volumen. Diese TF ist über die Intensitätswerte der beiden Datensätze definiert. Im gradientenbasierte Ansatz wird für jedes Volumen eine eigene TF spezifiziert und somit zwei separate Volumenvisualisierungen erzeugt, welche anschließend zu einer gemeinsamen multimodalen Visualisierung fusioniert werden können. Der gradientenbasierte Ansatz erlaubt eine interaktive Änderung der Transformation, welche die Registrierung der Datensätze beschreibt. Mit dem implementierten Volumenvisualisierungssystem wurden erfolgreich multimodale Visualisierungen erzeugt mit denen die Kommunikation von Diagnoseergebnissen erleichtert wird. Weiterhin konnten die Registrierungsergebnisse anhand der Visualisierungen überprüft werden.

Show publication details

Dold, Christian; Tingelhoff, Kathrin; Wesarg, Stefan; Firle, Evelyn

LV-Segmentation and Rotation Analysis Based on MRI Data

2005

Boenick, U. (Ed.) et al.: Medical Physics Vol. 1 : Proceedings of the Jointly Held Congresses ICMP and BMT. Berlin: Schiele & Schön, 2005. (Biomedizinische Technik 50, Suppl. 1), pp. 70-71

International Conference of Medical Physics of the International Organization for Medical Physics (IOMP) <14, 2005, Nuremberg, Germany>

Heart diseases often imply cardiac motion anomalies, so a number of heart diseases can be diagnosed by analysing cardiac volume and motion during one cardiac cycle. We propose a method for extracting complex left ventricle motion which consists of different components as a sinking motion caused by respiration, contraction and reverse rotation at cardiac basis and apex. The reverse rotation is determined by following landmarks over one cardiac cycle. An algorithm to detect automatically landmarks in CT and MRI data for complete heart motion analysis is the long term objective.

Show publication details

Schröer, Torsten; Sakas, Georgios (Betreuer); Wesarg, Stefan (Betreuer)

Segmentierung von Strukturen mit unscharfen Rändern in medizinischen Bilddaten: Implementierung, Validierung und Erweiterung des Region Growing Verfahrens nach Hojjatoleslami und Kittler

2005

Heilbronn, FH, Diplomarbeit, 2005

Gegenstand der Arbeit bildet ein 2D Region Growing Verfahren, welches sich durch die Behandlung von Objekten mit unscharfen Rändern auszeichnet, allerdings mit zunehmender Größe des zu segmentierenden Gebiets für den täglichen Einsatz zu langsam wird. Zunächst wurde daher ein 2D Prototyp auf Basis des Insight Segmentation and Registration Toolkit (ITK) erstellt, um das Verfahren, wie es in der Literatur beschrieben wird, zu validieren. Mit dem Prototypen konnten viel versprechende Ergebnisse, allerdings in unakzeptabler Zeit, produziert werden. Im folgenden wurde daher versucht, das Verfahren zu beschleunigen und auf einen 3D Algorithmus zu erweitern. Hierbei wurden unterschiedliche Datenstrukturen und Algorithmen eingesetzt und die Auswirkung auf die Geschwindigkeit des Algorithmus untersucht. Des weiteren werden zwei Multi Resolution Ansätze vorgestellt, die das Verfahren so weit beschleunigen, dass ein praktischer Einsatz auch für große Datensätze möglich wird. Da eine Beschränkung auf eine Modalität oder einen konkreten Anwendungsfall nicht gegeben war, wurde das Verfahren an unterschiedlichen Modalitäten getestet. Unter anderem wurden Computertomograhie-, Magnetzresonanz- und Ultraschalldaten verschiedener Körperregion ausgewählt, das Region Growing durchgeführt und die Ergebnisse qualitativ beurteilt bzw. mit Standard Region Growing Verfahren verglichen. Zusätzlich wurde ein Modell zur quantitativen Analyse des Algorithmus entwickelt. Das Modell wurde in Tests auf allen implementierten Varianten des Region Growings angewandt. Die Ergebnisse wurden verglichen und beurteilt. Es konnte gezeigt werden, dass alle 3D-Erweiterungen des Algorithmus auf unterschiedlichen medizinischen Bilddaten einsatzfähig sind. Dabei sind besonders die einfache Bedienung im Vergleich zu Standard Region Growing Verfahren (hoher Grad an Automatisierung) und die gute Behandlung von Kanten mit Gauß-verteilten Grauwertverfäufen hervorzuheben.

Show publication details

Wesarg, Stefan

Supporting the TECAB Grafting Through CT Based Analysis of Coronary Arteries

2005

Frangi, Alejandro F. (Ed.) et al.: Functional Imaging and Modeling of the Heart. Proceedings. Berlin, Heidelberg, New York: Springer, 2005. (Lecture Notes in Computer Science (LNCS) 3504), pp. 133-142

International Workshop, Functional Imaging and Modeling of the Heart (FIMH) <3, 2005, Barcelona, Spain>

Calcified coronary arteries can cause severe cardiac problems and may provoke an infarction of the heart's wall. An established treatment method is the bypass operation. The usage of a telemanipulation system allows for the execution of that operation as a totally endoscopic coronary artery bypass (TECAB) grafting. This relatively new method narrows the surgeon's view and does not permit the palpation of the vessel in order to detect calcifications (hard plaques). A planning based on contrast enhanced, cardiac CT data sets can compensate for that problem. This work presents analysis methods for coronary arteries. Hard plaques are detected using a tracking-based vessel segmentation technique. In addition, the vessel's neighborhood is analyzed in order to decide whether it is surrounded by tissue or fat, or if it is freely accessible for the surgeon's instruments. Furthermore, well adapted methods for the visualization of these analysis results are presented.

Show publication details

Tingelhoff, Kathrin; Dold, Christian (Betreuer); Wesarg, Stefan (Betreuer)

Tracking von Herzbewegungen aus Tomographiebildern

2005

Koblenz/Landau, Univ., Diplomarbeit, 2005

Herzerkrankungen äußern sich in vielen Fällen durch eine Veränderung der Herzbewegung. Sofern sich die Bewegung bestimmen lässt, kann man diese für die Diagnose von Herzerkrankungen verwenden. In dieser Diplomarbeit wurde die Herzbewegung aus CT- und MR-Bilddaten bestimmt, damit unterscheidet sich das eigene von den meisten bisher publizierten Verfahren, in denen datensatzspezifische Algorithmen entwickelt wurden. Auf der Grundlage einer ausführlichen Literaturrecherche wurden drei Konzepte entwickelt von denen eins im Rahmen dieser Arbeit implementiert und getestet wurde. Die berechneten Bewegungsparameter sind: Längsachsenkontraktion, globale Sack- und Kippbewegung des Herzens und eine Drehbewegung, die im apikalen und im basalen Bereich des Herzens entgegengesetzt ist. Die Bewegungsparameter wurden durch einen semiautomatischen Algorithmus bestimmt. Zur Berechnung der Längsachsenkontraktion und der Sack- und Kippbewegung wurde eine linke Ventrikel-Achse approximiert, deren Bewegung während eines Herzzyklus bestimmt und verfolgt wurde. Zur Bestimmung der Drehbewegung werden Landmarken händisch markiert, aus denen eine 3D Bewegung abgeleitet werden kann. Das entwickelte Verfahren wurde in die Medical Imaging Platform, ein in der Abteilung für Cognitive Computing and Medical Imaging entwickeltes Software-Toolkit integriert. Für die Implementierung wurden die open source Bibliotheken VTK, ITK, wxWidgets und boost verwendet. Das entwickelte Verfahren wurde mit CT- und MR-Bilddaten getestet und ausgewertet und es konnte gezeigt werden, dass durch das Verfahren medizinisch relevante Ergebnisse erzielt werden.

Show publication details

Schosser, Andreas; Wesarg, Stefan (Betreuer)

Visualisierung medizinischer Bilddaten unter Verwendung von Transferfunktionen für Farbe und Transparenz

2005

Gießen-Friedberg, FH, Diplomarbeit, 2005

Die Darstellung der mit Hilfe bildgebender Verfahren wie der Magnetresonanztomographie oder der Computertomographie gewonnenen medizinischen Daten erfolgt üblicherweise durch Grauwertbilder. Da beim Sehen nicht nur Helligkeitsempfindungen, sondern auch Farbton und Farbsättigung unterschieden werden, ist das farbliche Auflösungsvermögen des menschlichen Auges deutlich höher als die Auflösung unterscheidbarer Grauwert- bzw. Helligkeitseindrücke. Derzeitige Bestrebungen in der medizinischen Bildverarbeitung versuchen daher, sinnvolle Zuordnungen von Farbwerten zu den medizinischen Bilddaten zu finden. Im Rahmen dieser Arbeit konnten für verschiedene akquirierte Bilddaten medizinisch sinnvolle farbliche Darstellungen definiert werden. Dazu wurde in Anlehnung an die kommerzielle Anwendung VolView Volume Visualization eine interaktive Benutzerschnittstelle entwickelt, die eine flexible und intuitive Definition von Transferfunktionen erlaubt. Mit Hilfe der realisierten Visualisierungssteuerungskomponenten wird ein allgemeines Vorgehen für eine experimentelle Bestimmung medizinisch sinnvoller Farb- und Opazitätszuordnungen ermittelt und beschrieben. Ausgehend von den verwendeten bildgebenden Verfahren werden für verschiedene Körperbereiche individuelle Farbzuordnungen bestimmt und an konkrete medizinische Problemstellungen angepasst. Wie an verschiedenen Beispielen gezeigt wird, ist dadurch eine verbesserte Darstellung bestimmter Erkrankungen, Verletzungen und anderer Anomalien möglich.

Show publication details

Wesarg, Stefan; Firle, Evelyn; Schwald, Bernd; Seibert, Helmut; Zogal, Pawel; Röddiger, Sandra

Accuracy of Needle Implantation in Brachytherapy Using a Medical AR System - a Phantom Study

2004

The International Society for Optical Engineering (SPIE): Medical Imaging 2004: Visualization, Image-guided Procedures, and Display. Bellingham: SPIE Press, 2004. (Proceedings of SPIE 5367), pp. 341-352

SPIE Medical Imaging Symposium <2004, San Diego, CA, USA>

Brachytherapy is the treatment method of choice for patients with a tumor relapse after a radiation therapy with external beams or tumors in regions with sensitive surrounding organs-at-risk, e. g. prostate tumors. The standard needle implantation procedure in brachytherapy uses pre-operatively acquired image data displayed as slices on a monitor beneath the operation table. Since this information allows only a rough orientation for the surgeon, the position of the needles has to be verified repeatedly during the intervention. Within the project Medarpa a transparent display being the core component of a medical Augmented Reality (AR) system has been developed. There, pre-operatively acquired image data is displayed together with the position of the tracked instrument allowing a navigated implantation of the brachytherapy needles. The surgeon is enabled to see the anatomical information as well as the virtual instrument in front of the operation area. Thus, the Medarpa system serves as 'window into the patient'. This paper deals with the results of first clinical trials of the system. Phantoms have been used for evaluating the achieved accuracy of the needle implantation. This has been done by comparing the output of the system (instrument positions relative to the phantom) with the real positions of the needles measured by means of a verification CT scan.

Show publication details

Wesarg, Stefan; Seitel, Mathias; Firle, Evelyn; Dold, Christian

AHA Conform Visualization of Conventionally Acquired Cardiac CT Data Using the Toolkits itk and vtk

2004

Lemke, Heinz U. (Ed.) et al.: Computer Assisted Radiology and Surgery. Proceedings : CARS 2004. Amsterdam: Elsevier, 2004. (International Congress Series 1268), pp. 1096-1101

Computer Assisted Radiology and Surgery (CARS) <18, 2004, Chicago, USA>

Several modalities like CT, MRI, or U/S are used for the imaging of the myocardium and the adjacent cavity. The American Heart Association (AHA) has published recommenda- tions for the heart's orientation in cross-sectional slices for a better comparability of data from di erent imaging modalities. We present and compare techniques for the generation of such conform slices from conventionally acquired cardiac CT data sets by employing the toolkits vtk and itk for visualization as well as image processing purposes.

Show publication details

Firle, Evelyn; Wesarg, Stefan; Dold, Christian

Beschleunigte automatische CT/PET-Registrierung basierend auf partiellem Volumen-Matching und Mutual Informations

2004

Tolxdorff, Thomas (Ed.) et al.: Bildverarbeitung für die Medizin 2004. Proceedings : Algorithmen, Systeme, Anwendungen. Heidelberg: Springer, 2004. (Informatik aktuell), pp. 473-477

Workshop Bildverarbeitung für die Medizin (BVM) <2004, Berlin, Germany>

Die Berücksichtigung mehrerer tomographischer Bildgebungsmodalitäten nimmt eine zunehmend bedeutendere Rolle im Bereich der Diagnose und Therapieplanung ein. Hierdurch können vermehrt die Stärken der einzelnen Modalitäten genutzt werden. Insbesondere im Bereich der Krebsbehandlung bietet die Kombination von CT - zur Darstellung der Knochenstruktur -- und PET -- zur Visualisierung der funktionellen Information - etliche Vorteile. Das statistische Maß der Mutual Information ist eine Möglichkeit zur Bestimmung der für die Registrierung benötigten Transformation. Die hier vorgestellte Entwicklung beschäftigt sich mit dem Problem rigider Registrierung von CT- und PET-Volumina basierend auf Mutual Information. Es wird ein neuer Ansatz zur Beschleunigung des Matching-Prozesses unter gleichzeitiger Erhaltung der Genauigkeit und Robustheit dieser Methode vorgestellt.

Show publication details

Firle, Evelyn; Wesarg, Stefan; Dold, Christian

Fast CT/PET Registration Based on Partial Volume Matching

2004

Lemke, Heinz U. (Ed.) et al.: Computer Assisted Radiology and Surgery 2004. Proceedings : CARS 2004. Amsterdam: Elsevier, 2004. (International Congress Series 1268), pp. 31-36

Computer Assisted Radiology and Surgery (CARS) <18, 2004, Chicago, USA>

Combining CT representing bone structures and PET visualizing functional information reveals many advantages in the area of cancer diagnosis and treatment. Here, fully automatic registration is desired but very time-consuming when considering big datasets. We present a technique to register CT and PET datasets based on the "Mutual Information" criterion. To overcome the speed problem, we developed a new approach for the matching based on partial volumes introducing a "3D cross model".

Show publication details

Schwald, Bernd; Seibert, Helmut; Schnaider, Michael; Wesarg, Stefan; Röddiger, Sandra; Dogan, Selami

Implementation and Evaluation of an Augmented Reality System Supporting Minimal Invasive Interventions

2004

Workshop AMI-ARCS 2004. Online Proceedings : Augmented Environments for Medical Imaging [online]. [cited 4 February 2005] Available from: http://ami2004.loria.fr/, 2004, pp. 41-48

Workshop AMI-ARCS <2004, Rennes, France>

Minimal invasive surgery and interventions promise a number of advantages for the patient. The biggest one is a reduced trauma for the patient, at the cost of a limited view for the physician. The project MEDARPA offers Augmented Reality (AR) technology to enhance the real view of the surgeon with pre-operatively acquired medical data from 3D imaging modalities. During the last three years, a group of clinical, medical engineering and research partners has designed and implemented a prototype of an AR-enhanced navigation support system for minimal invasive interventions. It comprises of a transparent display device 'AR window', a hybrid tracking system, navigation support and a volume rendering system for medical 3D images. The system has been designed for a variety of medical applications. For the proof of concept cardiosurgery, radio oncology and pneumology scenarios have been considered. This paper gives an overview of the system design and its evaluation.

Show publication details

Wesarg, Stefan; Firle, Evelyn

Localizing Calcifications in Cardiac CT Data Sets Using a New Vessel Segmentation Approach

2004

Society for Computer Applications in Radiology (SCAR): SCAR 2004 : Extended Abstracts. Great Falls, Virginia: Society for Computer Applications in Radiology, 2004, pp. 14-16

Annual Meeting of the Society for Computer Applications in Radiology (SCAR) <21, 2004, Vancouver, BC, Canada>

The new generation of multislice CT scanners allows for the acquisition of high-resolution images of the heart. In cardiac surgery this image data is often the basis for interventional methods like the bypass grafting. In case of executing those interventions as minimal invasive procedures, it is mandatory to have a detailed knowledge about the position of calcifications in the coronary arteries. This work presents a method for localizing these calcifications by employing a newly developed vessel segmentation approach. That extraction technique has been developed for and tested with contrast-enhanced CT data sets of the heart. The algorithm provides enough information to compute the vessel's diameter along the extracted segment. An approach for automatically detecting calcified regions based on this information is presented.

Show publication details

Firle, Evelyn; Wesarg, Stefan; Dold, Christian

Mutual Information Based Registration for Ultrasound and CT Datasets

2004

The International Society for Optical Engineering (SPIE): Medical Imaging 2004: Image Processing. Part 2. Bellingham: SPIE Press, 2004. (Proceedings of SPIE 5370), pp. 1130-1138

SPIE Medical Imaging Symposium <2004, San Diego, CA, USA>

In many applications for minimal invasive surgery the acquisition of intra-operative medical images is helpful if not absolutely necessary. Especially for Brachytherapy imaging is critically important to the safe delivery of the therapy. Modern computed tomography (CT) and magnetic resonance (MR) scanners allow minimal invasive procedures to be performed under direct imaging guidance. However, conventional scanners do not have realtime imaging capability and are expensive technologies requiring a special facility. Ultrasound (U/S) is a much cheaper and one of the most .exible imaging modalities. It can be moved to the application room as required and the physician sees what is happening as it occurs. Nevertheless it may be easier to interpret these 3D intra-operative U/S images if they are used in combination with less noisier preoperative data such as CT. The purpose of our current investigation is to develop a registration tool for automatically combining pre-operative CT volumes with intra-operatively acquired 3D U/S datasets. The applied alignment procedure is based on the information theoretic approach of maximizing the mutual information of two arbitrary datasets from di.erent modalities. Since the CT datasets include a much bigger .eld of view we introduced a bounding box to narrow down the region of interest within the CT dataset. We conducted a phantom experiment using a CIRS Model 53 U/S Prostate Training Phantom to evaluate the feasibility and accuracy of the proposed method.

Show publication details

Wesarg, Stefan; Firle, Evelyn

Segmentation of Vessels: The Corkscrew Algorithm

2004

The International Society for Optical Engineering (SPIE): Medical Imaging 2004: Image Processing. Part 3. Bellingham: SPIE Press, 2004. (Proceedings of SPIE 5370), pp. 1609-1620

SPIE Medical Imaging Symposium <2004, San Diego, CA, USA>

Medical imaging is nowadays much more than only providing data for diagnosis. It also links 'classical' diagnosis to modern forms of treatment such as image guided surgery. Those systems require the identification of organs, anatomical regions of the human body etc., i. e. the segmentation of structures from medical data sets. The algorithms used for these segmentation tasks strongly depend on the object to be segmented. One structure which plays an important role in surgery planning are vessels that are found everywhere in the human body. Several approaches for their extraction already exist. However, there is no general one which is suitable for all types of data or all sorts of vascular structures. This work presents a new algorithm for the segmentation of vessels. It can be classified as a skeleton-based approach working on 3D data sets, and has been designed for a reliable segmentation of coronary arteries. The algorithm is a semi-automatic extraction technique requirering the definition of the start and end the point of the (centerline) path to be found. A first estimation of the vessel's centerline is calculated and then corrected iteratively by detecting the vessel's border perpendicular to the centerline. We used contrast enhanced CT data sets of the thorax for testing our approach. Coronary arteries have been extracted from the data sets using the 'corkscrew algorithm' presented in this work. The segmentation turned out to be robust even if moderate breathing artifacts were present in the data sets.

Show publication details

Tadge, Tonio; Wesarg, Stefan (Betreuer)

Segmentierung und Analyse von Herzgefäßen in CT-Daten

2004

Darmstadt, FH, Master Thesis, 2004

Jedes Jahr erleiden Millionen Menschen in den industrialisierten Nationen einen Herzinfarkt. Die Ursache hierfür ist oft eine Verengung (Stenose) der Herzkranzgefäße. Deshalb ist die Vermessung und Abschätzung der Geometrie der Herzkranzgefäße eine sehr wichtige Aufgabe bei der Diagnose von Erkrankungen. Es gab zwar schon längere Zeit die Möglichkeit, 3D-Aufnahmen zu erzeugen mit Hilfe der konventionellen Computertomographie (CT), aber erst die Erhöhung der räumlichen Auflösung auf 0.5 mm Kantenlänge pro Voxel und der Einsatz von Kontrastmitteln hat diese für die Kardiologie interessant gemacht. In dieser Arbeit wurden verschiedene Verfahren zur Segmentierung und Analyse von Herzkranzgefäßen untersucht und in die Medical Imaging Platform (MIP) des Fraunhofer Institut für Graphische Datenverarbeitung Darmstadt integriert. Dabei wurden nach einer Literaturrecherche drei Verfahren zur Segmentierung ausgewählt und implementiert und zusammen mit einem in dieser Arbeit entwickelten Verfahren hinsichtlich Robustheit, Geschwindigkeit, Genauigkeit und Benutzerfreundlichkeit getestet und verglichen. Dabei konnten die Verfahren Confidence Connected (Flächenwachstumsverfahren) und Freezing Fast Marching (Modellbasiertes Verfahren) wegen ihrer Ungenauigkeit bei der Segmentierung nicht überzeugen. Dagegen lieferten die Algorithmen Resample Verdonck (Verfolgungsverfahren [tracking based]) und der neu entwickelte Vessel Cutter brauchbare Ergebnisse. Für die Analyse von Herzkranzgefäßen wurde ein die Krümmungen der Gefäßwand betrachtendes Verfahren für die Klassifizierung von Stenosen und Aneurysmen ausgewählt und implementiert. Da die Ergebnisse dieser Analyse eher zweifelhaft waren, wurden zusätzlich zwei Charakteristika zur Klassifizierung verwendet. Das war der Querschnitt des Gefäßes und ein Gradienten-basiertes Kriterium im Resample Verdonck- und Vessel Cutter- Verfahren. Zusammenfassend läßt sich sagen, dass der Einsatz der in dieser Arbeit beschriebenen Verfahren auf die dreidimensionalen, hochaufgelösten CT-Daten wertvolle und neue Erkenntisse für die Diagnose der relativ kleinen und schwer zu erkennenden Herzkranzgefäße gebracht hat.

Show publication details

Firle, Evelyn; Wesarg, Stefan; Dold, Christian

Registrierung und Visualisierung von 3D U/S und CT Datensätzen der Prostata

2003

Handels, H. (Ed.) et al.: Bildverarbeitung für die Medizin 2003. Proceedings : Algorithmen, Systeme, Anwendungen. Heidelberg: Springer, 2003. (Informatik aktuell), pp. 26-30

Workshop Bildverarbeitung für die Medizin (BVM) <2003, Erlangen, Germany>

Brachytherapie ist eine Strahlentherapie, welche u. a. mit Hochenergie-Strahlenquellen in Hohlnadeln, die in den Körper des Patienten eingestochen werden, durchgeführt wird. Die präzise Konturierung des zu bestrahlenden Gewebes, sowie die genaue Plazierung der Hohlnadeln an den Positionen, welche durch das "Pre-Planing" vorgegeben werden, sind hierbei wichtige Arbeitsschritte. Bisher basiert die Behandlung des Prostatakarzionoms mittels Brachytherapie vornehmlich auf CT Aufnahmen, welche aber keine Echtzeit-Visualisierung während der Implantation der Katheter zulassen. Sind sowohl CT als auch 3D U/S Aufnahmen vorhanden, können diese registriert und fusioniert werden, um somit die Vorteile beider Modalitäten zu nutzen. Im folgenden werden die Untersuchungen zur Registrierung sowie Möglichkeiten zur Evaluierung dargestellt.

Show publication details

Firle, Evelyn; Wesarg, Stefan; Karangelis, Grigorios; Dold, Christian

Validation od 3D Ultrasound - CT Registration of Prostate Images

2003

Sonka, Milan (Ed.) et al.: Medical Imaging 2003: Image Processing. Proceedings. Bellingham: SPIE Press, 2003. (Proceedings of SPIE 5032), pp. 354-362

SPIE Medical Imaging Symposium <2003, San Diego, CA, USA>

All over the world 20% of men are expected to develop prostate cancer sometime in his life. In addition to surgery - being the traditional treatment for cancer - the radiation treatment is getting more popular. The most interesting radiation treatment regarding prostate cancer is Brachytherapy radiation procedure. For the safe delivery of that therapy imaging is critically important. In several cases where a CT device is available a combination of the information provided by CT and 3D Ultrasound (U/S) images o.ers advantages in recognizing the borders of the lesion and delineating the region of treatment. For these applications the CT and U/S scans should be registered and fused in a multi-modal dataset. Purpose of the present development is a registration tool (registration, fusion and validation) for available CT volumes with 3D U/S images of the same anatomical region, i.e. the prostate. The combination of these two imaging modalities interlinks the advantages of the high-resolution CT imaging and low cost real-time U/S imaging and o.ers a multi-modality imaging environment for further target and anatomy elineation. This tool has been integrated into the visualization software "InViVo" which has been developed over several years in Fraunhofer IGD in Darmstadt.

Show publication details

Schnaider, Michael; Seibert, Helmut; Schwald, Bernd; Weller, Tanja; Wesarg, Stefan; Zogal, P.

Medarpa - Ein Augmented Reality System für Minimal-Invasive Interventionen

2002

Bundesministerium für Bildung und Forschung (BMBF): Internationale Statustagung "Virtuelle und Erweiterte Realität". [cited 20 July 2005] Available from: http://informatiksysteme.pt-it.de/vr-ar-2/tagungscd.html, 2002, 9 pp.

Internationale Statustagung Virtuelle und Erweiterte Realität (VR/AR) <2, 2002, Leipzig, Germany>

Minimal-invasive Operationstechniken erfreuen sich in den letzten Jahren einer zunehmenden Akzeptanz und Verbreitung. Sie bieten dem Patienten offensichtliche Vorteile wie ein geringes Trauma, kürze Behandlungszeiten, etc. Gleichzeitig steigen mit diesen Techniken aber auch die Herausforderungen an die Chirurgen, die sich nun mit einer stark eingeschränkten optischen Wahrnehmung des Operationsbereiches konfrontiert sehen. In diesem Beitrag wird ein neuartiger Ansatz für die Computer-unterstützte Chirurgie vorgestellt, der die Techniken der Visualisierung und der erweiterten Realität (AR - Augmented Reality) mit einem neuartigen Ausgabegerät, dem AR-Fenster, kombiniert. Das Forschungsprojekt Medarpa beschäftigt sich mit der Entwicklung eines medizinischen AR-Systems, das ein im Projekt entwickeltes halb-transparentes Display nutzt, welches - über den Körper des Patienten geschwenkt - dem Chirurgen eine Sicht in den Patientenkörper ermöglicht. Es werden auf dem halb-transparenten Display im Vorfeld aufgenommene medizinische 3D-Bildmodalitäten eingeblendet, Anatomien dargestellt und in den Patientenkörper eingeführte Operationsinstrumente visualisiert. Der Chirurg erhält eine erweitere Darstellung der Realität, die ihm in seiner Aufgabe unterstützt. Es wurde ein erstes prototypisches Gesamtsystem entwickelt, dessen Funktionsüberprüfung in ersten Testläufen bei den klinischen Partnern bereits erfolgreich durchgeführt werden konnte. Dabei wurde bereits während der ersten Entwicklungsphasen auf die Berücksichtigung der Anforderungen eines realen klinischen Einsatzes Wert gelegt.

Show publication details

Wesarg, Stefan; Ebert, Matthias; Bortfeld, Thomas

Parker Weights Revisited

2002

Medical Physics, Vol.29 (2002), 3, pp. 372-378

The short-scan case in fan-beam computed tomography requires the introduction of a weighting function to handle redundant data. Parker introduced such a weighting function for a scan over pi plus the opening angle of the fan. In this article we derive a general class of weighting functions for arbitrary scan angles between pi plus fan angle and 2pi (over-scan). These weighting functions lead to mathematically exact reconstructions in the continuous case. Parker weights are a special case of a weighting function that belongs to this class. It will be shown that Parker weights are not generally the best choice in terms of noise reduction, especially when there is considerable overscan. We derive a new weighting function that has a value of 0.5 for most of the redundant data and is smooth at the boundaries.

Show publication details

Firle, Evelyn; Chen, Wei; Wesarg, Stefan

Registration of 3D U/S and CT Images of the Prostate

2002

Lemke, Heinz U. (Ed.) et al.: Computer Assisted Radiology and Surgery 2002. Proceedings. Berlin, Heidelberg, New York: Springer, 2002, pp. 527-532

Computer Assisted Radiology and Surgery (CARS) <16, 2002, Paris, France>

Radiotherapy is a rapidly growing cancer treatment technique. In brachytherapy - one radiotherapy treatment technique - pre- and post-planning is usually carried out using CT imaging. As CT scanners cannot easily be moved from one operation room to an other and as CT does not have real-time imaging capability, alternative imaging modalities are needed to realize the vision of image guided surgery. Ultrasound (U/S) is such an alternative imaging modality. For the comparison of U/S and CT image fusion is very useful. After volume segmentation (see ref. 1) the volumes have to be registered, and afterwards the fused volume can be displayed. In this paper we investigate the registration part. We present two different approaches for the 3D registration of CT and U/S volumes. They are compared regarding accuracy and speed of calculation. The resulting fused volumes are visualized using the "InViVo" software where the registration routines have been integrated into.