Model Construction of Nonrigid Biological Objects from Images

Being able to accurately model the deformation of biological objects is vital to the success of many multidisciplinary studies such as computer-aided surgery simulation and planning, image-based medical diagnosis, and motion-based human identification etc. Among various modeling techniques, physical model becomes increasingly popular because the physical and mathematical rigors upon which the model is established can ensure physically sound motion predictions. It has been widely recognized that material properties assigned to the physical model could have a significant impact on its prediction accuracy. In this talk we will explore a number of applications of such modeling for burn scar evaluation and breast cancer detection to facial expression analysis and biometrics.