Wednesday October 11, 2006, 8:30-9:30
Convention Hall No. 1, Level 2

Human Intention Modeling and Interactive Computer Vision

Abstract:

For many years, computer vision and robotics researchers have worked hard chasing the illusive goals such as "can the robot find a boy in the scene" or "can your vision system automatically segment the cat from the background". These tasks require a lot of prior knowledge and contextual information, and perhaps more importantly, understanding of human intention. How to model human intention into vision and robotic systems is, however, very challenging and can only be solved through human-computer interaction. In this talk, we propose that many difficult vision tasks can be solved with interactive vision systems, by combining powerful and real-time vision techniques with intuitive and clever user interfaces. We will show two interactive vision systems we developed recently, Lazy Snapping (Siggraph 2004) and Image Completion (Siggraph 2005). Lazy Snapping cuts out an object from a picture using graph cut, while Image Completion recovers unknown region in a picture with belief propagation. A key element in designing such interactive systems is how we model the user's intention using conditional probability (context) and likelihood associated with user interactions. Given how ill-posed most image understanding problems are, it is proposed that interactive computer vision is the paradigm we should focus today's vision research on where the key is the understanding and modeling of human intention.

Biography:

Dr. Harry Shum brings his extensive research skills, excellent management capabilities, and outstanding academic background to Microsoft Research Asia (MSRA), Microsoft's basic research arm in Asia-Pacific, where he serves as the Managing Director overseeing and directing research activities and university relations.

Dr. Shum is on the Editorial Board of IEEE Transaction's Pattern Analysis and Machine Intelligence (PAMI) and International Journal of Computer Vision (IJCV) and served as the General co-Chair for the Tenth International Conference on Computer Vision (ICCV 2005 Beijing). He has published more than 100 papers on computer vision, computer graphics, pattern recognition, statistical learning and robotics and has been awarded more than 20 U.S. patents. Dr. Shum is a Fellow of the Institute of Electrical and Electronic Engineers (IEEE).

Dr. Shum joined Microsoft in 1996, first as a researcher in the Vision Technology Group of Microsoft Research in Redmond, Washington where he worked on computer vision and computer graphics. He then moved to Beijing, China in January 1999 to help start Microsoft Research China (later renamed as Microsoft Research Asia). He was subsequently promoted to a Research Manager, Senior Researcher, Assistant Managing Director, and Managing Director of MSRA.

Prior to joining Microsoft, Harry Shum was a Senior Research Scientist with the San Jose based start-up Realspace and completed internships at Apple Computers' Interactive Media Lab and Digital Equipment Corporation's Cambridge Research Lab.

Dr. Shum received a Ph.D. in Robotics from the School of Computer Science at Carnegie Mellon University. In his spare time, Harry enjoys watching a good football match, rooting for the Pittsburgh Steelers, and spending quality time with his family.