Trellis Codes Facing Video Object

With the development of information society, people are increasingly demanding multi-media communications. And the huge amount of information of multimedia communication, multimedia communication costs are also very large. Video information is very rich and a huge amount of data communication information in multimedia communications, people on the visual requirements of the video information is increasing. Therefore, we must increase the ratio of data compression of video information to improve the communication channel bandwidth in order to achieve a lower communication cost multimedia communications. So we introduce a video image coding technology, first video data be compressed then data storage and transmission, video image compression encoding has become an important task of the video information processing.Grid coding in wireless communications get a good application, this combine channel coding with modulation technology, we can get a good coding gain. Also the grid encoding successfully applied to video encoding applications, DT grid code is the use of Delaunay triangle mesh of the fine the geometric characteristics, the video image is divided into DT adaptive grid, the current MPEG-4 standard has adopted DT grid code and has been well studied in video coding, and DT grid code is gradually improving development and the implement application, to get better results.To achieve the compression coding of video information, motion estimation is a process can not be ignored. In video sequence among of the adjacent image frames with strong time correlation, motion estimation technology could remove the time redundancy in such video information to achieve the purpose of compressed video data. Because of the importance of motion estimation, motion estimation algorithm and its improved algorithm is endless, all to get more better compression ratio and efficiency.This article change the pristine motion estimation block-matching methods, want to improve blocking effects for block-matching and a huge amount of computation. Firstly, this article described video image frame by DT grid, adaptively divide the video image, according to changes of image material to adaptively change the grid. In where the rich image information, grids is more and smaller; in where the smooth image information, grids is less and bigger. Secondly, nodes of the DT grid carry out motion estimation using radiation transformation. This article presents a new improved affine transform method, different from previous affine transformation with the grid-based motion estimation algorithms, not only can effectively model more complicated motion such as camera rotation, zoom, or object shearing, When multiple objects passing each other, can reconstruct the scene, and reduce the adverse impact of the noise and light changes. It can improve the accuracy of motion estimation, has a good improvement relative to block-effect caused by block matching algorithm and the complex DTM campaign algorithm.