Dissertation
Dissertation > Industrial Technology > Radio electronics, telecommunications technology > Communicate > Image communication, multimedia communication > Image coding

Research on Key Technologies in Multiview Video Codec

Author YanZuo
Tutor ZhouJun
School Shanghai Jiaotong University
Course Signal and Information Processing
Keywords Depth estimation Depth Based Region extraction Disparity Compensation GPGPU
CLC TN919.81
Type Master's thesis
Year 2010
Downloads 204
Quotes 4
Download Dissertation

As the development of information technology, network technology and multimedia technology, the rapid development of the visual enjoyment of people’s demands have been rising. Over the past few years, the video system developed basically by increasing the number of pixels to improve image sharpness. However, due to the human eye visual resolution limit, the future image system may develop basically by increasing the number of video’s views to provide real-world multiview three-dimensional senses. Multiview video is the future development direction.Accurately obtain the scene depth information is the foundation of high-performance three dimensional video encoding and image processing. This paper proposed an improved belief propagation algorithm for stereo matching. Based on the assumption that the disparity field is continuous, traditional methods regard the disparity field as a Markov network that transmits two-way information. But in the occluded area, disparity is not continuous. So we propose a new method. First, we use the cross-check technology based on the initial disparity to detect the occluded area. Second, we regard the disparity map as a mixed network of Markov field and Bayes filed. Then the occluded area does not transmit information to the non-occluded area so as to reduce the computational cost of disparity matching. We use the standard test images to evaluate our algorithm. The result shows that the proposed method has a high accuracy and efficiency.Multiview video need store and transmit more than one point of view scenes video, data volume is a traditional two-dimensional video several times. Studying the efficient compression coding technology for the practical application of multiview is very important. In general, all kinds of the three-dimensional video coding scheme will use disparity estimation technologies to reduce redundancy between views, this paper presents an improved a fast algorithms which can improve disparity- estimation efficiency. On the other hand, this paper studies the depth based prospects for the region extraction algorithm and carried out three-dimensional video foreground and background encoding attempts. At lower bit-rate environment, foreground and background encoding can improve image quality in the prospected region.On the other hand, due to three-dimensional multiview video decoder needs to synchronize multiple viewpoint video real-time decoding, decoding performance is thus a constraint to the practical application of multiview video decoder. This paper proposed a method based on CPU and GPU mixed decoder solution, which improved the multiview video software decoding efficiency.

Related Dissertations
More Dissertations