Dissertation > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Pattern Recognition and devices > Image recognition device

Researches on Depth Compression in 3d Video

Author JiMaoSheng
Tutor LiHouQiang
School University of Science and Technology of China
Course Signal and Information Processing
Keywords 3D video coding Scene change Layered depth video technology Perspective synthesis
CLC TP391.41
Type Master's thesis
Year 2010
Downloads 177
Quotes 0
Download Dissertation

3D (Three Dimension 3D) video emerging forms of media, and is in the stage of vigorous development. Compared to traditional two-dimensional video, 3D video can provide the depth information of the object observed, allowing the user to generate a three-dimensional visual experience, some of the three-dimensional technology allows users to enjoy the show free choice point of view, to switch perspective. On the application of 3D technologies are expected to market is very extensive, including 3D movies, 3D video conferencing, 3D satellite imagery, human-computer interaction perspective selection function. However, three-dimensional video in addition to the normal color of the information contained herein, the additional depth video, which led to a huge amount of data, the ability to greatly compressed 3D video data will become the main limitation of this media technology can be a large range of applications condition. At the same time, three-dimensional video show some new characteristics, such as the motion correlation between the depth information and color information, does not have the texture information, etc. Another example is the depth information, which provides a new solution for 3D video compression algorithm developed high efficiency ideas, which will also promote the application of three-dimensional video technology market. Therefore study the three-dimensional video compression technology not only has a wide range of applications and has some theoretical significance. 3D video contains multi-angle color information and depth information, this form of media organizations, we study two key technologies of the three-dimensional video-depth information processing, the main work and innovation lies in: 1. Propose a new The use the compression depth information to improve the efficiency of the three-dimensional video coding method of color information and depth information is to build the basic elements of the three-dimensional scene, and neither is dispensable. Studies have shown that the human eye is sensitive to the degree of depth in the scene change is much smaller than the sensitivity of the color, which is not allowed during this time to provide in-depth information. This paper presents a scene change-oriented three-dimensional video compression method, based on the existing virtual view generation technology: layered depth video technology (Layered Depth video, LDV) and video compression technology, can significantly improve the overall three-dimensional video compression performance. The corresponding subjective tests show that this method does not degrade the subjective visual quality of the 3D video. (2) study the impact of the depth video quality synthetic virtual view quality explore the depth video quality perspective synthetic effect can effectively guide the 3D video encoding bit rate allocation color information and depth information. Users are only interested in a virtual composite video perspective synthesis process used in the color information and depth information are not visible at this time is not the need for both to meet quality optimal synthesis best virtual video is worth open to question, there are already studies have shown that a suitable proportion allocate bandwidth to different video can give the user a good visual experience, adding more bandwidth corresponding depth information does not give the user more screen to enjoy. The article continues through experiments proved that the depth of the low-resolution video for viewing angle Synthesis and does not significantly reduce the the synthetic Virtual Perspective mass of, i.e., the low quality of the depth information is available.

Related Dissertations
More Dissertations