A Sparse Coding Algorithm For Fractal Image Compression Based On Coefficient Of Variation

Fractal image compression coding occupies an important position in the field of image compression because of its high compression ratio and high quality of image reconstruction,but only using image compression fractal technology cannot take advantage,and the hybrid coding technology combining fractal and other advanced technologies has a better development prospect.Fractal image compression coding needs matching the best match from the vast amounts of code blocks,and the process requires a lot of time,leads to the basic fractal algorithm with high computational complexity and long encoding time,high compression ratio and high reconstruction quality problem.It cannot be satisfactory.Based on the basic problems of fractal image compression algorithm,improve the quality of reconstructed image and improve the encoding speed point of view,mainly from the following aspects:(1)Aiming at the problems of high complexity and lengthy coding time of fractal image coding algorithm,orthogonal sparse coding and texture feature extraction are proposed to represent image blocks.Firstly,the orthogonal sparse transformation of gray level improves the image reconstruction quality and shortens the decoding time.Secondly,the correlation coefficient matrix measures the characteristics of variation coefficients between range blocks and domain blocks to reduce redundancy and coding time.Simulation results show that the proposed method has better image reconstruction quality and faster encoding speed compared with the traditional fractal image coding algorithm.(2)In view of the time-consuming problem in the process of image block matching,the paper proposes to simplify the search stage in the process of image block matching.In this chapter,the method of extracting the chi-square feature of image block is adopted..The chi-square feature of image block is extracted,and the complexity of image block search is reduced by using the statistical property of chi-square feature.Then,the similarity measurement of chi-square feature of image block is carried out.Finally,sparse coding is used to reduce the redundancy of feature and save coding time.Experimental results show that the proposed algorithm has significant advantages over the basic fractal image compression algorithm.(3)using the statistical features of the image block itself,this chapter defines the new characteristics of image block,first calculate the variance of image block,variance normalized processing,then the variance of the processed do integral,using the geometric meaning of variance integral image block area-variance characteristics,finally extracted area-variance characteristics as orthogonal fractal sparse coding matching the best image block matching feature vectors search phase.Experimental results show that the algorithm can improve the quality of reconstructed images and shorten the coding time.