Research On Segmentation Of Moving Objects In Video Sequence

With advances in communication and information processing technologies, video-driven applications show a great ability of flexibility and extensibility. Visual communication is the fastest growing vehicle for information. Many new digital applications and services are emerging such as digital TV, teleconference, videophone, and image-based interactive multimedia. These diversified applications and services with a large amount of data demand more advanced digital signal processing techniques for efficient storage and transmission, accurate analysis, and flexible manipulation. Video object segmentation is the base of the mentioned techniques.Video object segmentation aims to partition an image sequence into moving objects and to track the evolution of the moving objects along the time axis. Many applications related to video compression and transmission ,and pattern recognition rely on video object segmentation Video object segmentation techniques are also important tools for content-based video coding and manipulation, and interactive multimedia applications. Video object segmentation usually divides the contents of a video frame into semantic regions that can be dealt as objects. These semantically segmented objects can be coded so that object-based manipulation of video content can be realized in interactive multimedia applications.This paper first present the development of video object segmentation techniques, and state the recent status of this area. The basic segmentation methods were introduced and summed up into several classes. In the following, this paper introduced the block matching algorithm which has a close relationship with motion estimation and according to the experiment result, we choose the best fast matching algorithm to fulfill our work. Then, we advanced two modified motion segmentation algorithm: one is based on region motion vector , this algorithm can improve the performance of the segmentation mask; another is modified semi-interactive segmentation algorithm based on seeded region growing. In order to adapt to the non-translational movement, we mended the dissimilarity measure of the algorithm, and get better results.