- Vita
- Publications
Danwu Chen received his B.Sc. in 2011 from Department of Electronic and Information Engineering, South China University of Technology, China, and his M.Sc. in 2013 from Department of Electrical Engineering and Computer Science, National Taiwan University of Science and Technology, Taiwan.
From early 2014 to mid-2019, he has worked as an algorithm engineer at China Software Center of ASML (the worldwide leading supplier to semiconductor industry), with 50% responsibility for C/C++ software development related to computational geometry and another 50% responsibility for machine learning (especially designing and improving deep learning models (in terms of both accuracy and runtime) on the company's image data in order to optimize mask designs for manufacturing integrated circuits).
In mid-2019, he has joined the Fraunhofer Institute for Computer Graphics Research IGD, Darmstadt, as an EU researcher. His research direction is machine learning for computer graphics in 3D printing.
Deep Learning Models for Optically Characterizing 3D Printers
2021
Optics Express
Multi-material 3D printers are able to create material arrangements possessing various optical properties. To reproduce these properties, an optical printer model that accurately predicts optical properties from the printer’s control values (tonals) is crucial. We present two deep learning-based models and training strategies for optically characterizing 3D printers that achieve both high accuracy with a moderate number of required training samples. The first one is a Pure Deep Learning (PDL) model that is essentially a black-box without any physical ground and the second one is a Deep-Learning-Linearized Cellular Neugebauer (DLLCN) model that uses deep-learning to multidimensionally linearize the tonal-value-space of a cellular Neugebauer model. We test the models on two six-material polyjetting 3D printers to predict both reflectances and translucency. Results show that both models can achieve accuracies sufficient for most applications with much fewer training prints compared to a regular cellular Neugebauer model.