Abstract
I describe physics-based models of human motion, with applications to computer animation and 3D person tracking. I begin by surveying relevant principles of motion from the biomechanics literature. I then describe a representation of motion that incorporates passive dynamical elements, relative muscle preferences, and optimality principles. I then describe Nonlinear Inverse Optimization, a novel learning algorithm that can be used to learn these physical models of motion from data. Once learned, these models can then be synthesize new animation. I then describe a tracking algorithm for on-line estimation of 3D human motion from video sequences that makes use of a physics-based prior model of motion.

Joint work with Marcus Brubaker, David Fleet, Zoran Popovic, C. Karen Liu

Bio
Aaron Hertzmann is an Associate Professor of Computer Science at University of Toronto. He received a BA in Computer Science and Art & Art History from Rice University in 1996, and an MS and PhD in Computer Science from New York University in 1998 and 2001, respectively. In the past, he has worked at University of Washington, Microsoft Research, Mitsubishi Electric Research Lab, Interval Research Corporation and NEC Research Institute. His awards include the MIT TR100 (2004), an Ontario Early Researcher Award (2005), a Sloan Foundation Fellowship (2006), a Microsoft New Faculty Fellowship (2006), and a UofT CS teaching award (2008). His research interests include computer vision, computer graphics, and machine learning.