Automatic Moving Objects Segmentation for Head-Shoulder Video Sequence

摘要

Video Object (VO) segmentation, which requires the decomposition of sequences into semantically meaningful physical objects, is one of key problem in content-based image coding system. In this paper, we present an automatic moving object tracking and extracting algorithm. The algorithm employs the high correlation between the edge of frame difference and the edge of gray level image. Experiment results demonstrate the performance of our method.