The Research Of Video Image Compression Based On The Self-adaptive Partitioning

The image compression is an important task in the field of image processing. The aim of the image compression is to cut down the redundant data of the image without damaging the valid information of the image. At present, during the arithmetic regularly used in the field of image compressing arithmetic, the distinct character is that it averagely segments the image, and then carries out the translation and coding to every block.But every image is different to each other and every one has the different distribution about smooth region and marginal region. From the viewpoint of compression, compression with big dimension to the big smooth region and with small dimension to the small marginal region can improve efficiency and quality prominently. Herein this article brings forward the compressive arithmetic of self-adaptive blocking. First of all, segments the image with the method of clustering based on force function and different tabs are given to the different region. Secondly, considered to the tabs, quadtree partitioning is used to segment the image block with the size of 32 and the arithmetic ends when the size of the block is 4. At last, the segmenting block is coding with different dimension, including DCT transform based on digital, quantification, journey coding, difference coding, Huffman coding and segmenting information coding. The segmentation method adopted by this article is efficient and capable;meanwhile, the fast DCT transform is designed and it boosts the efficient of the whole arithmetic.Based on above idea, this article puts forward self-adaptive estimation to motion vectors and makes the abecedarian discussion. The foundation is that movement is small to the adjacent frame with the exception of scene changing;the still region has a large proportion;and the strong relativity exists in the adjacent frame. The method is getting rid of the still regions based on the diff-frame. At the same time, self-adaptive segmentation is used to the diff-frame in order to estimate the move vectors.From the achievement of arithmetic and the analysis of results, compared to the traditional compression such as JPEG, this arithmetic improves the compressiveefficiency and quality of rebuilding image. Besides, it also has some advantage to the self-adaptive methods appeared and more efficient. Meanwhile, the self-adaptive estimation of move vectors has betterment to the traditional methods.