一种基于分块的分形图像压缩方法

分形编码具有高压缩比的特点,但在编码过程中其匹配搜索时间开销巨大.提出一种基于分块的分形图像压缩方法:依据人眼视觉特性在PSNR变化不大的情况下,图像主观质量感受不明显的特点,将图像在空域分割为若干相对较小的块,对感兴趣区可采取一分为四的方法进一步减小块的大小,然后对每个块分别进行分形编码,并重构.结果表明,在PSNR略有降低的情况下,编码的匹配搜索时间大幅减少,同时重构图像的主观质量并没有明显降低.

A fractal image compression method based on block-dividing

For the convenience of store and transmission,it is necessary to conduct compression encoding for digital image.Fractal image coding has the characteristics of high compression,but the match searching time in the process of encoding is huge.In this paper a fractal image coding method based on block-dividing is proposed.The human vision system is not sensitive to the subjective quality of the reconstructed image in case of a small PSNR change.The original image is divided into small blocks in space domain directly,and to the interesting parts the blocks can be divided further by one-to-four.Then the fractal image coding to each blocks is carried out,and the reconstruction is also conducted ateach block respectively.The experimental results demonstrate that the match searching time can be reduced substantially and the subject image quality remain unchanged basically,meanwhile the PSNR being decreased slightly.