A Robust Algorithm for Video Based Human Motion Tracking

In this paper, we present a robust algorithm to capture rapid human motion with self-occlusion. Instead of predicting the position of each human feature, the interest-region of full body is estimated. Then candidate features are extracted through the overall search in the interest-region. To establish the correspondence between candidate features and actual features, an adaptive Bayes classifier is constructed based on the time-varied models of feature attributions. At last, a hierarchical human feature model is adopted to verify and accomplish the feature correspondence. To improve the efficiency, we propose a multiresolution search strategy: the initial candidate feature set is estimated at the low resolution image and successively refined at higher resolution levels. The experiment demonstrates the effectiveness of our algorithm.