Event Location: NSH 1507
Bio: Simon Lucey is an Associate Research Professor within the Robotics Institute at Carnegie Mellon University. Prior to this he was an Australian Research Council (ARC) Future Fellow (2009-2013) and a Principal Research Scientist at Australia’s national science agency the Commonwealth Scientific and Industrial Research Organisation (CSIRO). His research is motivated from a passion for discovering the “why?” behind “how?” with respect to core problems in Artificial Intelligence, Computer Vision and Machine Learning. Simon leads the CI2CV laboratory, one of the leading groups in the world for face and body analysis through computer vision. He received his Ph.D. in 2003 from the Queensland University of Technology, Australia. To his credit he has over 90 peer reviewed publications. He is currently an Associate Editor for IEEE Transactions on Affective Computing, and has served as Area Chair for conferences like ICPR (2012), CVPR (2014 & 2015), ACCV (2014), and F&G (2015). He has also organized and chaired numerous workshops, tutorials and summer schools in Computer Vision and Machine Learning.

Abstract: This is a joint work with Y. Zhu, D. Huang, and F. de la Torre. The task of estimating complex non-rigid 3D motion through a monocular camera is of increasing interest to the wider scientific community. Assuming one has the 2D point tracks of the non-rigid object in question, the vision com- munity refers to this problem as Non-Rigid Structure from Motion (NRSfM). In this paper we make two contributions. First, we demonstrate empirically that the current state of the art approach to NRSfM (i.e. Dai et al.) exhibits poor reconstruction performance on complex motion (i.e motions involving a sequence of primitive actions such as walk, sit and stand involving a human object). Second, we propose that this limitation can be circumvented by modeling com- plex motion as a union of subspaces. This does not naturally occur in Dai et al.’s approach which instead makes a less compact summation of subspaces assumption. Experiments on both synthetic and real videos illustrate the benefits of our approach for the complex nonrigid motion analysis.