Motion-Based Hierarchical Active Contour Model for Deformable Object Tracking

Abstract

This paper proposed a novel scheme for combined contour extraction and deformable object tracking. In order to track fast moving objects, we first add the motion estimation term to the energy function of the conventional snake. Then, a hierarchical approach using wavelet analysis is applied. Although the proposed wavelet-based method can track objects with large motion, the proposed method requires less computational load than the conventional one. By using a training procedure, the proposed method overcomes occlusion problems and local minima due to strong edges in the background. The proposed algorithm has been tested for various images including a sequence of human motion to demonstrate the improved performance of object tracking.

This work was supported by Korean Ministry of Science and Technology under the National Research Laboratory Project, by Korean Ministry of Information and Communication under the Chung-Ang University HNRC-ITRC program, and by the Korea Research Foundation Grant funded by Korean Government (MOEHRD)(R08-2004-000-10626-0).