The Fractal Coding In The Wavelet Domain

With the development of computer science, image compression becomes more and more important in the communication system and multimedia system, and it affects our lives, production and national defense greatly. Therefore, people have shown great concern of image compression. The international compression standard, having DCT as the core has formed. These years, A fractal image coding based on wavelet transform is showing its advantage, in which a wavelet transform has heen adopted by JPEG-2000. So this image coding has the tendency of replacing the image coding of having DCT as the core. The author in this papers have done some research of image compression, it contains three Parts.Firstly, having a whole cognition of image compression coding, the thesis analyses the advantages of the adoption of DCT by many international standard.In the second part: The emphasis of this part lies in discussing Subband coding, Wavelet transform coding and Fractal coding, whose coding functions are better than The coding of having DCT as the core. Especially the latter two coding, since 1990 have attracted people抯 attention widely. Doing the research of the internal system of these two codings and finding the best connection between them will play great role in the development of image compression.In the third parts: Based on Davis?self-Quantization of subtrees, the fractal coding in the wavelet domain put forward Orientaiton Wavelet-subtree Fractal coding, This coding encodes individually towards wavelet-subtree, at the same time, Davis put forward the image coding concerning the disadvantage existing in Rinaldo and Calvagno抯 fractal predictive case. They adopted different ways towards wavelet coefficient singly and could not make use of spatial partial individually features such as cross-scale fringe existed widely. Ahthough Davis?iiSQS image coding has achieved higher coding efficiency, but he hasn抰 thought about the differences between relativity degree and feature of the wavelet subtrees in different directions, according to the differences of the relativity among them, different scan methods have been adopted in coding. Moreover, because of the weakness of taking too much time in Fractal coding, zerotree wavelet has been used flexibly in the process of coding, which deleting a lot of predictable and unimportant coefficient, at the same time making them equal to zero, thus the coding time has been shortened. In a word, the method put toward in this paper makes the speed faster, improves PSNR and compression ration in a certain degree so it is an effective image coding.