Template protection – also known as biometric encryption or cancelable biometrics – can convert biometric data into a secure, non-invertible form. The system generates reference data records from which users cannot retrieve any biometric information. Stored biometric identities can be protected efficiently with the help of the template protection method. This increases user acceptance of biometric systems considerably.
Flyer 'Template Protection'
Multi-biometric processes combine various biometric features such as fingerprint, iris and vein patterns in order to make a more reliable and flexible decision whether it is actually the authorized person accessing a home, smartphone or car. If you combine varoius processes and merge the resulting data by means of "information fusion", it becomes easier to verify or identify the right person. There is a series of technologies for biometric end devices but only few of them have established products in the marketplace so far. The most well-known applied processes are fingerprint, iris and 2D face recognition. Fraunhofer IGD further develops known processes, works on new methods an evaluates systems available in the marketplace. This includes in particular 3D face recognition, signature recognition and ear recognition. Each individual process is very secure per se although the used devices have a fault tolerance depending, on the manufacturer and may also be overcome depending on their quality.
Biometrics provide a source to establish unique identity of an individual. The unique identity can be in the form of a national ID card, electronic passport or a smart token used for everyday life activities. The unique identity can be used for ID check in large scale identification system, namely Automated Biometric Identification System (ABIS), Automated Fingerprint Identification System (AFIS). In cases where ID check is to performed in a system of a vast deployment then the ID check becomes a time-taking process and may no longer be useful where quick decisions need to be made. Indexing is the name of a process where similar data in the database can be grouped together. Because of indexing, the ID check process can be conducted only on the groups that have a resemblance to the test template. Competence center Identification and Biometrics (IDB) of Fraunhofer IGD has developed an indexing approach that can be used to index databases of face templates. The approach can be used to index multi-biometric databases as well.
Biometrics on Smartcards
EC and credit cards are becoming more secure by means of biometrics. Until now, customers have been paying by PIN or signature. PINs can be stolen and signature images can be forged. We use people’s signature dynamics as biometric characteristics so that every card is only assigned to the rightful owner. Signatures are biometric characteristics like fingerprints, iris or vein patterns. Everybody always has his signature ready without having to learn it by heart, like a PIN. The smartcard compares the biometric features of the card user and the rightful card holder directly on the card. During this on-card comparison, the biometric reference data remain securely stored on the smartcard, leaving it at no time.
Flyer 'Convenient Security due to biometrics'
Operators of biometric systems are confronted with the problem of a lack of security measures which protect against forgery. There are ways to cheat biometric systems using dummies and the existence of such possibilities affects confidence in biometric systems and in biometrics as a whole. The "Identification and Biometrics" department has developed image analysis methods to deal with this problem. These include biometric methods and systems which can detect if the subject is alive and if it displays the biometric characteristics which match those of a living person, rather than a copy - such as a photograph or a mask.