Research On Related Technique Of Video Coding For Stereoscopic Images

Stereoscopic video technique which can produce three-dimensional effect is becoming the focus of study of the future multi-media technique. The application of stereoscopic video technique, however, is limited because of the enormous data. This paper is in reference to disparity estimation, information concealment and error concealment for stereoscopic images.Disparity estimation which is an effective method of getting rid of space redundance can compress stereoscopic video data efficiently. It has been studied a lot at home and abroad. In this paper the constraint conditions of disparity estimation are first analyzed in terms of the parallel camera system. And then a new scheme of disparity estimation is proposed in terms of reducing computing load. Experimental results show that the scheme is effective in accelerating computing speed in comparison with the full search method.Currently, the watermark schemes of video are far immature comparing to those of still images. The characteristics of motion estimation in H.264 are first studied and then a new watermark scheme for stereoscopic video is proposed by changing the motion vectors got by the half pixel motion estimation and search range of the quarter pixel motion vectors. The scheme can be expanded to detect the transmission error. It has been proved effective by many experiments.Error concealment is used to recover the video images damaged by erroneous channels. The technique achieves this using the correspondence of space and time and adding additional information to the video. After analyzing the schemes of single view video, this paper puts forth a new error concealment scheme based on multiple block patterns. The scheme makes full use of the motion characteristics among different blocks and disparity correspondence of the two views. Experimental results show that the scheme is very effective.