Bit Allocation And Rate Control Technologies For Multi-view Video Coding

Multi-view video is a very important developing direction in video technologies in future. It includes multiple video sequences captured by placing multiple cameras at the same time, but different locations and from the same scene and includes the depth information of scene, which can provide the third dimension and the reality to people. Along with the rapid development of multimedia communication technologies, there is a very broad application prospect in digital entertainment field, such as free viewpoint video, free viewpoint television, 3DTV and so on. However, because of the increased number of cameras, the amount of video data is also rapidly increasing, making the compression coding and transmission of multi-view video data become very difficult. Rate control is one of the key factor influencing the multi-view video transmission quality. The study subject of this paper is the research on the rate control algorithm for multi-view video based on the statistics of the multi-view video coding. The main works of the paper are as follows:First of all, by reading lots of correlative literatures of multi-view video coding technology and rate control algorithm, this paper introduces the research actuality and development tendency of multi-view video coding and rate control technology and describes its basic principle. the paper intensively researches on multi-view video coding and the related reference software model JMVC, analyzes the facts of the rate control accuracy, such as rate distortion model, header information bits prediction model and the initial quantisation parameter. Based on the statistics of mult-view video coding, a hierarchical rate distortion, a a hierarchical header information bits prediction model and adaptive initial quantisation parameter are proposed in this paper.Second, a bit allocation strategy based on consistent view quality for multi-view video coding is proposed. the bit allocation contains three layers: GGOP layer, GOP layer and frame layer. In order to obtain the consistent view quality, firstly, a bit allocation model based on Lagrange optimum algorithm is proposed. Secondly, taking into account the encoding statistical characteristics of different view types, a view weighting factor is introduced, and it will help improving on the precision of bit allocation among views.Finally, aiming at the multi-view video coding structure, a rate control algorithm for multi-view video coding is proposed in this paper, the rate control algorithm contains three layers: GGOP layer, GOP layer and frame layer. The objective of the GGOP layer rate control is to estimate of the target bits for each GGOP and determine the initial quantisation parameter for the coding first frame in the sequence. The objective of the GOP layer rate control is to estimate of the target bits for each GOP and determine the initial quantisation parameter for the coding first frame in each GOP. The objective of the frame layer rate control is to estimate of the target bits for each frame and determine the quantisation parameter for each non-key frame. At last, this paper summarizes the work of this subject and shows the expectation of further hard work.