Motion capture is a prevalent technique for capturing and analyzing human articulations. However, some marker positions are often missing due to occlusions or ambiguities. Interpolation based methods only work well when the markers are missing only for a very short time. We propose a data-driven piecewise linear modeling approach to missing marker estimation. We model motion sequences of a training set with hierarchy of low-dimensional local linear models characterized by the principal components. For a new sequence with missing markers, we use the classifier to identify the most likely model for each frame and then recover the missing markers by finding the least squares solutions based on the known markers and the principal components of its associated model. We demonstrate in the experiments that after offline training, our method can efficiently recover the missing markers and its ability to generalize over a variety of motions from multiple subjects.
Guodong Liu, Leonard McMillan. Estimation of Missing Markers in Human Motion Capture. To appear in Proc. of the 2006 Pacific Graphics, Taipei, Taiwan, October 2006.
Paper in PDF format (399 Kb)
Video in AVI (48M, zipped)
Presentation in PPT (2M)