A Study Of Image Coding Based On Fractal

Nowadays, the multimedia technology develops rapidly. And how to store, transferinformation efficiently and conveniently is the most important problem we have to deal with.Picture as the maximum media of data volume, implementing high function of compressionbased on the premise of acceptable restore status. Using less bits to express the original pictureinformation is the key to solute this problem. At present, in the field of image compression,there has been formed a series of international standards on the foundation of DCT、huffman,such as JPEG, MPFG, h.263, etc. However scholars have never stopped to find better methodsfor image compression. Because of the potential for high compression ratio,resolutionindependence and rapid decoding characteristics, fractal coding become more and morepopular in this field.Fractal coding is based on the principle of the local image similarity, using a group ofapproximate constant compression affine transformation express image, the only fixed point isreconstruction images. In the decoding, the transformation of the original image fractal codeswhich combined by the compression of the parameters, the decoded image is infiniteapproximated by the contractive transformation denoted in the fractal on any initial image.This paper make the fractal coding algorithm as the research object, and how to improvethe encoding speed of fractal coding and quality of decoding image as the main researchpurposes. Combined with the statistical properties of actual image, on the basis of lots ofresearch achievements, I did following aspects of the work:1. This paper also proposes a fast decoding scheme, let the mean of range blocks instead ofthe gray offset, make it independent of the whole search matching process, enhanced thedecoding speed. Realize2iteration convergences of reconstruction image.2. To simplify the scaling factor and make the calculation of it before8transformation.soas to reduced the amount of calculation. Then we analysed the distribution of scaling factorsand discussed it influence to the coding speed. Based on this, proposes a new algorithm toreduce the searhing code based on the simplified scaling factors, which achieves a good effect.3. Verify the rationality of the correlation coefficient and standard deviation to improvethe matching ratios of range block and domain block, then set a threshold, when the standarddeviation of range block, make a mean of range block to instead of it.the same to the domainblock,if the standard deviation of domain less than the threshold,take it out of the searhing code.After this to size correlation coefficient as another judgment standard, whenever the rangeblock and domain block meet the conditions will be matching computation. The algorithm optimizatied the search scope, reduce the calculation of coding and make the encoding speedimproved.