Research On Low Complexity Depth Map Coding Algorithm Based On HEVC

In the age of digital information, with the rapid development of technology, the traditional 2D video has been unable to meet people’s visual experience. Multi-view video plus depth(MVD) has become a central issue in the field of video research. As an important part of MVD, depth map stores the depth information of the scene and can be used to synthesis virtual viewpoints at the receiving end to reduce the amount of data for storage and transmission. Currently the common method to code depth map is the HEVC standard. Different from traditional video, depth video typically has a lot of viewpoints which means more coding complexity. At the same time, HEVC improves the coding complexity greatly. These factors have restricted the application of depth map. Since depth map has the characteristics of smooth regions and sharp edges, its properties of coding are different form traditional videos and can be used to reduce the coding complexity of depth map coding based on HEVC.This paper firstly analyzed the characteristics of depth map and introduced the latest video coding standard, HEVC, and some main techniques like coding unit, intra prediction and inter prediction. Secondly, it introduced depth map coding based on HEVC and analyzed the source of coding complexity. Then it made a research of the best mode for depth map intra prediction and proposed a fast intra prediction algorithm based on fast edge detection and fast CU partition. At last, the characteristics of inter prediction for depth map were studied and a fast inter prediction algorithm based on CU similarity was proposed. After simulation, experiments showed that the two proposed algorithm can effectively reduce the encoding time in the premise of guaranteeing the quality of the compression and coding.