Exemplar-based human contour tracking

We propose an exemplar-based tracking framework for human contour tracking. The exemplars, i.e. the contour representatives, are used to construct a contour ensemble. The main task of contour ensemble is to generate contours to fill in the gaps in-between in the test sequences, and to supply the dynamics for updating the target contour by fast contour query. As a result, a normal dynamic Bayesian network is only used to infer the location and the size of the target contour. The effectiveness of the proposed method is tested by many human motion sequences.