Bayesian Reconstruction of 3D Human Motion from Single-Camera Video
|MERL Report: ||TR99-37: Nicholas R. Howe, Michael E. Leventon, William T. Freeman
Advances in Neural Information Processing Systems 12}, edited by S. A. Solla, T. K. Leen, and K-R. Muller, 2000
Three-dimensional motion capture for human subjects is underde- termined when the input is limited to a single camera, due to the inherent 3D ambiguity of 2D video. We present a system that re- constructs the 3D motion of human subjects from single-camera video, relying on prior knowledge about human motion, learned from training data, to resolve those ambiguities. After initializa- tion in 2D, the tracking and 3D reconstruction is automatic; we show results for several video sequences. The results show the power of treating 3D body tracking as an inference problem.