The Research And Improvement On Algorithm Of Fractal Image Compression Based On Irregular Area

The rapid development of computer technology and networks, especially that of multimediatechnology, makes us convenient and enjoy the sight. However, it also brings us the problem of how toreduce the cost of store and transmit a lot of image information at the same time. It needs to compress theimage information effectively basing on not affecting the visual perception. But there are manydisadvantages in the traditional coding methods, such as the serious block effect in the decoding image witha high compression ratio, etc. Therefore, many new coding schemes are proposed, among which the fractalcompression algorithm becomes one of the research hotspots because of its high compression ratio,resolution of independence and fast decoding. Whereas the automatic image fractal compression algorithmis not ideal enough in terms of the encoding time, compression ratio and effect. So, it has great theoreticalsignificance and practical value to implement a fast algorithm for fractal coding or improve the fractalcompression method.This paper first introduces the background and significance of the fractal compression as well as itscurrent research status, expounds the basic principles and the math theory of fractal compression, and stillthere is a brief introduction of the main improved algorithm of fractal compression together with theprinciple of the key technologies in irregular area segmentation encoding technology. Then, this article putsforward a PCNN-GRAY algorithm of fractal compression combined with the irregular regionsegmentation. The image blocks are classified according to the gray value of the binary image whichobtained by dividing the original image with the method of irregular region segmentation and that of theoriginal one. On the premise of the quality of decoding image, the algorithm not only reduces the searchrange a lot but also improves the coding speed. Then a fractal compression algorithm based on imagecomposite feature classification is proposed, which is improved based on the PCNN-GRAYalgorithm.Experiments show that the algorithm can significantly reduce the coding time withoutsacrificing the quality of decoding. At last, to reduce the block effect arising from the proposed algorithm, this article has carried on the classification of overlap and smooth processing, which can effectivelyeliminate the block effect.