Research Of The Encoding Of Stereo Video Conference System

Because of the increasement of depth's information, three-dimensional video scene drew much attention as it can enhance the visual sense of authenticity and realism and become a very important development direction of video technology. But compared to 2D image and video system, three-dimensional video system needs to store and transport a very large amount of data, therefore, it is very necessary to research on high-speed, high-efficiency, high-quality compression methods.Video coding standard is not only a common basis of single-channel video compression, but also the basic of three-dimensional video compression method, three-dimensional video compression encoding is based on ordinary one-channel video encoding and can be achieved based on its improvements. This article focuses on the improvement of efficiency of the current video compression model to prepare for the realization of stereo video conference system.H.264/AVC video coding standard is currently the latest international image coding standard, and it has better performance than any other standard before.H.264/AVC has a wide range of applications because of its high encoding efficiency, good network compatibility. HEVC is the phototype of the next generation of video compression coding standard, asking it to have better coding efficiency. HEVC and H.264 have a similar basic framework, but not rigidly adhere to the form of H.264/AVC standard. Not difficult to predict, as H.264/AVC, HEVC's superior encoding performance, and the advantages of the all aspects of H.264/AVC video coding applications showing at present, they will be able to lead a breakthrough of the three-dimensional video on the storage and transmission, and guide the three-dimensional video and all kinds of stereo video applications to achieve greater improvements.This paper will give a brief introduction of three-dimensional video and stereo video conference system, then based on the understanding of the key technology H.264, chapter three and chapter four proposed two improvement of the current video encoding algorithms. One of the improvements is based on the block-based coding model to propose a fast intra prediction method, and the other is an adaptive reference frame prediction algorithm according to the object's motion synthesis that based on the principle of multiple reference frame model.