| Distributed video coding is a new type video coding scheme, using intra-frame coding to reduce encode complexity and using inter-frame coding in decoding. This coding mode will transfer the coding complexity from coding end to decoded end, which is fit for wireless video sensor and mobile phone applications which have low power equipment. So it's attracted by many organizations and researchers in recent years. Multi-view technology is a kind of three-dimensional video technology, which can get scene and 3-d information more vividly and accurately reproduce in film,3-d scenes, remote control and virtual reality. Now, multi-view Video Coding (MVC) has become an active research. This article focuses on the realization of distributed video coding(DVC), image store and transmit protection by DVC, multi-view coding based on distributed video coding. The main content includes the following aspects:First, realized the basic framework of distributed video coding. The software is based on c/c++ and OpenCV library and convenient to transplant to various systems. With flexible architecture, support pixel domain, transform domain distributed video coding. Support multiple types of macro block and video size. Support several LDPCA codec. Support non-uniform quantification and reconstruction algorithm based on the side information. Support bitrate dynamic adjustment. Support the virtual viewpoint synthesis.Secondly, realized image store and transfer applications based on distributed video coding. Propose a set of different noise models under distributed coding scheme, and apply it to the picture of the preservation and transmission. Respectively analyse three different noise models:salt&pepper noise, the ink noise and quantization noise, and propose different decoding schemes according to their respective characteristics. Apply these schemes to enhanced security, robustness, and compression rate in image preservation and transmission.Finally, realized a multi-view coding scheme based on distributed video coding. Realized a viewpoint synthesis method based on depth chart, using multiple basic viewpoints to improve synthetic viewpoint image's quality, using shelter and hole filled technology to improve image quality further. Apply the enhanced synthetic viewpoint image as side information to build a new, low bit rate and high quality multi-view distributed video coding system. |
PDF Full Text Request