Research On Multiview Video Rate Control Algorithm

With the rapid developments of information, network and multimedia technologies, multiview video encoding technology, which can provide viewers with the benefits of stereo visual effect, become one of research highlights in the video domain in recent years and will be widely used in the fields of stereo video conferencing, stereoscopic television, remote robot control system, telemedicine, etc. However a large amount of video data with multiview video is one major obstacle for storage and transmission of multiview video data. Therefore, it is vital to compress stereo video efficiently and rate control in one of the key link. This works in this thesis are based on human stereo visual feature. We study the multiview video rate control algorithm in two ways of reducing the bit rate and improving the perspective.In the first part, we briefly introduce the background of stereo video and expound the digital video compressed encoding technology. And we review the status quo of research at home and abroad from two aspects of multiview video rate control and foreground and background segmentation encoding-based rate control as the technical background of the following part.Secondly, we mainly study the basis of multiview video encoding. By studying the multiview video encoding framework proposed by standardization organization at home and abroad, we present a multiview video encoding framework based on MPEG-2 and illustrate the position of rata control in it and the relation between rate control and other modules. Besides, we introduce the technology of motion estimation and motion compensation and the technology of disparity estimation and disparity compensation and point out that the application of these two technologies has a direct effect on rate control. Thirdly, we proposed a multiview video rate control scheme based on human stereo visual feature. We introduce the new quantitative measure for stereo video quality based on the feature. Then, we derive the frame level bit allocation scheme on the basis of this measure. Finally, we present a method of modifying the macroblock quantization parameters according to the quantization parameter of the neighboring macroblock on the purpose of eliminating block effect in stereo video. The experiment results demonstrate that the proposed scheme can improve the stereo video quality as well as obtain more stable rate.Finally, we proposed a rate control scheme which is appropriate for the foreground and background segmentation encoding. We derive the macroblock level bit allocation scheme on the basis of different priorities of foreground and background. Moreover, we present the adjust method for macroblock quantization parameters considering of the difference of foreground and background and blocking effects. The experiment results demonstrate that the proposed scheme can improve perspective of the stereo video as well as control the stability of output rate effectively.