Fractal Image Coding Method Based On Texture Feature

With the developing of multimedia technology and computer communication, digital image which has enormous data quantity restricts image communication. Effective encoding technologies to obliterate redundancy and retain the image quality are the focus of research. In this dissertation, from theoretical and practice viewpoint, a novel approach to compress algorithm, fractal image compression algorithm based on texture character is discussed, which is different from the traditional method. The kernel theory based on in this dissertation is called Iterated Function System (IFS). The main idea is to find the IFS which consists of a set of contractive affine transformations mainly based on fixed-point theorem and collage theorem, using block division method based on statistical and texture characteristics. Decoding process starts from any images which can recur to the original image by applying IFS. Therefore, only IFS codes need to be stored for coding file, which can achieve high compression ratio.The dissertation firstly goes to details of the descriptions of image texture features, an improved texture image retrieval method is proposed, combining statistical methods with structural ones naturally. It can be used to describe the gray distribution and the local detail structure in a textured image and has remedied deficiency of unitary method, moreover, it can describe image with less features. Be compared with other algorithm, this algorithm is more precision.Secondly, based on statistical characteristics of image, an improved fast fractal coding algorithm is proposed by classifying with domain blocks and range blocks, which make up for the fractal algorithm based on entropy. It will lose the best match block when the entropy value of the range lies between boundary areas, decreasing image quality. The algorithm avoids the lose of the best block by determining thresholdδ, and the image quality can be improved. Experimental results demonstrate that the proposed algorithm can considerably shorten the encoding time, while achieving the better decoded image quality as baseline fractal algorithm with full search.Finally, improved hybrid fractal coding algorithms based on texture features have been compared with other methods. Combining the image texture features with fractal coding is an effective method for improving fractal image coding in terms of quality. In the hybrid coding, variation function which could describe image texture feature is applied to fractal image coding. On this basis, a new texture character feature is suggested and it is proved that affine transformation can't change the value of it. The above analysis provides a sound theoretical basis for its application. Both theoretical analyses and experiments show that higher compression ratio and image quality can be achieved, and the method is feasible.