Research On Method Of H.264-based Stereoscopic Pictures Compression

With the development of economy and the improvement of living standard, the requirement for real sense of image becomes higher and higher. Stereo video technology has become one of the new developments in the field of vision. But compared to the 2D images, 3D video system has to store and transmit huge amounts of data. So, data compression is necessary in 3D system. Meanwhile, a high-speed, high-efficiency and high-quality compression algorithm is very necessary.H.264 video coding standard is the latest international standard for image coding. Its performance is much better than the previous standards. With the high coding efficiency and the good network compatibility, it must have a wide range of applications. Therefore, H.264-based stereoscopic pictures compression is feasible.This paper introduces the principle of stereoscopic vision and the key technologies of H.264. And we describe an image compression algorithm based on the H.264 for stereo video. For stereo video compression, the left view can be considered as main view, and can be compressed based on H.264. The right view can be compressed based on the left view using both motion-compensate prediction and disparity-compensate prediction. And with the inter prediction, intra prediction and other H.264 tools, we can get an efficient compression. This paper also introduces some key technologies of H.264, eg. multiple reference pictures for motion compensation, entropy coding and so on. It proves that these technologies are efficient by emulating.The simulation results show that the algorithm is efficient. It can access to higher compression ratio and higher quality of compressed images. Further more, it shows a good prospect of H.264 application in this field.