Nonparametric Density Estimation for Human Pose Tracking

The present paper considers the supplement of prior knowledge about joint angle configurations in the scope of 3-D human pose tracking. Training samples obtained from an industrial marker based tracking system are used for a nonparametric Parzen density estimation in the 12-dimensional joint configuration space. These learned probability densities constrain the image-driven joint angle estimates by drawing solutions towards familiar configurations. This prevents the method from producing unrealistic pose estimates due to unreliable image cues. Experiments on sequences with a human leg model reveal a considerably increased robustness, particularly in the presence of disturbed images and occlusions. In Pattern Recognition, Springer LNCS 4174, K. Franke et al. (Eds.), pp. 546-555, Berlin, Germany, Sep. 2006 c Springer-Verlag Berlin Heidelberg 2006

• Type: Conference
• Year: 2006
• Where: DAGM
• Authors: Thomas Brox, Bodo Rosenhahn, Uwe G. Kersting, Daniel Cremers
• Keywords: DAGM 2006, Image Processing, Joint Angle, Joint Angle Configurations, Joint Angle Estimates