基于分形的图像编码技术研究计算机应用技术专业论文.docxVIP

优秀毕业论文 精品参考文献资料 ABSTRACTTo ABSTRACT To effectively store and transmit the massive image data IS the important problem which the information age faces now．Because of its necessity and feasibility,the study on image encoding and decoding already becomes more and more active．Fractal image coding is a very promising compression technique．It has been paid great attention since its presence because of its novelty，high compression rata，resolution independence and rapid decoding．In this dissertation，some self-similar， fast fractal image encoding and decoding scheme are proposed，as follows： 1．A method that estimates image’S self-similarity is presented and the factor denoting character of image’S self-similarity is defined through studying the self-similarity model of image．Experimental results indicate that the bigger self-similarity parameters，the beRer image coding efficient． 2．Because existing fast encoding methods are often at the cost of reconstructed image quality，or Call only obtain poor speed—up ratio，this paper further proposed some methods which improved both quality and speed of fractal encoding：through combining the self-similarity coefficient and clustering algorithm，using the average of range blocks-variance as the center of initialized clustering and sorting self-similarity parameters，it reducing search time，optimizing the algorithm in the process of affine transformation with reducing the matching domains．Experiment indicates the method which this paper proposed accelerate fractal image encoding． 3．The method makes the best use of the theorem of resolution independent in fractal theory．It improves affine transformation decoding algorithm and combines affine transformation and iterating the average‘ of range blocks to improve the decoding algorithm．Experiment indicates the method not only improve the quality of reconstruct image， but also reduce the decoding time． KEY WORDS Image compression，fractal，Iterated Funciton Systems， self-similarity,Cluste