| Summary: | 碩士 === 國立中正大學 === 電機工程所 === 97 === Due to highly development of Internet and multi-media among these few years, many new types of applications appear. Such as, wireless PC camera, cellphone and multi-media network. Traditional video encoding, couldn’t meet the requirements anymore. They both execute complicated motion evaluation when encoding. Therefore, the whole costs are enhanced highly by this kind of complicated encoding way. To save more costs of system, lowering the complication from encoding site to decoding site becomes the most important issue: Distributed Video Coding.
By the principle made by Slepian-Wolf and Wyner-Ziv, we could design an effective video encoding structure. In the encoding structure from Wyner-Ziv, almost encoding technologies are based on Pixel-domain. And, the re-built data would be the side information or the boundary value around the side information when signal responded. But, the disadvantage of this way is that only single data of pixel would be considered, and correlation of area districts is ignored. To preserve the edge correlation and win the better video technology in the limited bandwidth, this essay is forwarded to add Dynamic Programming into re-built site.
Doing Dynamic Programming at re-built site makes the pixels which don’t consider near data originally preserve the edge correlation for the near pixels, so the quality would be improved after re-building. The result of simulated experiment shows that maximum 0.5dB unit could be enhanced after the decoding structure of Pixel-domain combined with the Dynamic Programming in the same bit rate.
|
|---|
