| Multi-view video, which can provide viewers with the benefits of added realism, selective viewing, and improved scene understanding, will be widely used in the fields of 3D TV, free-viewpoint TV, immersive videoconferencing, virtual reality etc. However a large amount of data is one major obstacle for using multi-view video is the large amount of data. A multi-fold increase in bandwidth over the existing single-view makes it extremely tough to transmit and store multi-view video data. This thesis mainly concerns the problems of highly efficient multi-view video coding (MVC). To achieve high compression efficiency, co-relation between the different views must be exploited in multi-view video coding scheme. Based on this scenario, we propose various efficient encoding schemes.Firstly, we proposed an H.264-based multi-view video coding scheme. It uses the advanced predictive method of H.264 to eliminate the spatial, temporal and inter-view co-relation in multi-view video. According to the characteristic of multi-view video, global disparity coding method is employed. An eight-parameter global disparity model is used and two global disparity coding modes are proposed. To decrease the coding bit number of motion vector (MV) and disparity vector (DV), an optimized MV&DV predicted method is proposed. It utilizes not only the MV&DV of casual neighboring blocks, but also that of corresponding blocks in the adjacent images. To support view-scalability, a layered disparity prediction structure is proposed. Using the layered disparity prediction structure, our MVC stream can be random accessed and decoded partly.Due to the distinct performance of cameras, there exists serious color-variation (including brightness variation) between the images of different views. These variations well impact the compression performance of multi-view video coding. To improve the coding efficiency of MVC, we investigate the method of color-variation compensation disparity prediction. Firstly we analyze and model color-variation based on the image formation theory. Then, various color-variation compensation methods are proposed according to the simplified models. They are global linear color-variation compensation, global non-linear color-variation compensation, local color-variation compensation and global-local-adaptive color-variation compensation. Experiment results shows these methods can greatly improve the performance of disparity prediction and the coding efficiency of MVC.Rate-distortion (R-D) analysis and rate control play an important role in video coding and communication systems, which are to prevent buffer malfunction and provide the highest possiblevideo quality under the constraints of rate and delay. In this thesis, rate control algorithms of both traditional single-view video coding and MVC are provided. Due to the high computational complexity of multi-pass variable bitrate (VBR) rate control algorithm, we proposed a low-complexity single-pass VBR rate control algorithm for real-time video coding application. It allocates the target bit number for each picture according to its complexity, and can achieve constant video quality. Although the rate control of video coding is deeply investigated, rate control of MVC is seldom concerned. In this thesis, we proposed a rate control algorithm for H.264-based MVC scheme. The algorithm employs an independent quarter rate-distortion model for each viewpoint video. Simulation results showed that it could accurately control the used bit number of each frame.
|
PDF Full Text Request