Video Image Analysis And Optimization Of Motion Predication

The amount of digital image is great and the amount of digital video is many times of that of digital image. The hugeness of data takes great burden of memory unit, communication channel or computer processing speed. It is unpractical to resolving this problem only just by increasing memory capacity ,channel bandwith or computing speed. That needs many money and not every user can take it. So, it is a good way to study the technique of data compression. It is one of key technique of video analysis and motion predication in video compression. So, the emphasis of this paper is optimization of motion predication based video analysis.First, I learned some corresponding context of the new video-compression standard of H.264. And presented an optimizing algorithm of macro-block mode choosing for P-frame by using analysis the motion of video frame in H.264 standard. This algorithm was proved to reduce coding time about to half of that before. To optimizing predication of I picture, I studied some methods of video analysis and emphasize the methods of morphologic and watersheds. I combined with these methods and the characters of video image to segment I picture in H.264.The watersheds method was modified and histgram regulation method was used to enhance video image.The result of experiment showed that it's good and effectively prevent over-segmentation.In this paper, I aimed at the special perpose of video-conference and emphasize the interesting area that is human face. My segmentation experiment of face in video sequence was proved to be good. At last I learned the theory of motion compensation in H.264 and combined with works in last steps to optimize the motion predicition in H.264.I emphasized the coding data and coding time of face area.The result of experiment shows that it effectively reduces coding time.However, I did't consider the optimization of coding data organization, so my algothm increases the amount of coding data and reduce the compression ratio. Improving compression ratio is my work in future.