The Research Of Image Compression Coding Based On Fractal

In the past few years the image compression coding method based on fractal theory has attracted more attention of domestic and foreign researchers. Through making the full use of self-similarity, the fractal image coding methods eliminate the jumbled image information and achieve the compression of image data. But the shortcomings of fractal image coding methods are that the encoding time is too long and the computational complexity is high. Against these shortcomings the following works were done in this paper.(1) A fast matching algorithm based on correlation coefficient was improved, which searched the range blocks of maximum correlation with domain block instead of searching the best matching domain blocks. In order to improve the encoding speed only the correlation coefficient in the primary and secondary on the diagonal elements in domains were computed. Experimental results showed that it reduce the encoding time with the guarantee of quality of recover image.(2) A kind of discrete cosine transform(DCT) coding algorithm were improved. In this method the coefficient of in left upper corner an the coefficient bigger than the given threshold were retained. The method reduced the image information loss. The experimental results showed that the recovery image quality was improved.(3) A algorithm of image compression based on the combination of fractal theory and DCT transform was put forward. By using fractal similarity larger image block compensated discontinuity of smaller blocks. The experiment result shows that this algorithm improved the recovery image quality and reduce the phenomenon of “block effect”.