3DTV Coding Based On Layered Depth Video

From2D to3D, is an important development direction of digital video systems. In the past more than ten years,3D display technology has made considerable progress in a variety of3D display technology matures and compete to launch the market products. Based on the coupling grating and2D flat panel displays from the auto-stereoscopic display technology is the most cost-effective at this stage bare as the3D display, which allows the viewer to get rid of the shackles of3D glasses, naturally perception of3D visual content, thus becoming the new ideal choice for generation of three-dimensional television system.In this paper, the research and development of the laboratory natural3D TV key technologies and systems, demand-oriented multi-view video in3D terminal reconstruction, focusing on the efficient expression of3D TV content and coding techniques. The main results include:(1) In the synthesis and analysis of existing international and national3D video coding standards and technologies based on the proposed layered depth video (LDV:Layered Depth Video) the natural3D TV system architecture is proposed and implemented based on multi-camera to capture video LDV generation algorithm, based on the The LDV multi-view virtual video reconstruction algorithm to study the characteristics of the LDV data and multi-view video reconstruction quality. The experimental results show that, with respect to the single-depth video, based on the the3D LDV expression program, can be an effective supplement to cover information, data increments of less significant reduction in the empty area in the multi-viewpoint video reconstruction, to improve the image quality.(2) Based on the H.264standard modifications and extensions, proposed and implemented a new LDV codec system. To block layer macroblock is divided into the the boundary macroblock and non-boundary macroblock two types, a series of new codec technology for border macro block, macroblock mode adaptive selection, polynomial fitting, decoding side exact boundary reconstruction, loop filter, adaptive control, and extend the standard syntax to support the new coding tools. Experimental results show that, relative to the LDV coding techniques have been proposed in the literature, the proposed scheme can effectively reduce the bit rate, the organization and layered scalable compressed stream, suitable for network transmission.