Stereo imaging in low bitrate video coding

A method for integrating stereo imaging principles into video coding is described along with the progress of the research up to now. Specifically, the basic structure of an m×384 kbit/s (m=1.5) video codec as recommended by the CCITT is described. The same structure also forms the basis of the work on recommendations for an n×64 kb/s (n=1.6) codec. Problems that could be encountered by using the structure for such low-bit-rate coding are discussed, including distortion introduced by the coding process and the results of the generation of incorrect motion vectors. The use of depth information in the coding process to overcome these problems is suggested. The depth information is extracted from the image by the application of stereo imaging principles. A discussion of the basic principles of stereo imaging is followed by two suggested methods for integrating stereo imaging into the video coding process. The first is an extension of the CCITT video coding structure, using the depth information to identify the problem areas so that special steps can be taken. In the second method the blockwise motion detection and image prediction of the CCITT video codec structure are replaced by a system based on a 3-D model of the moving object within a scene constructed from the depth information