基于人类视觉系统的彩色图像分形编码算法

在分析传统的四叉树分形编码算法的基础上,将人类视觉系统(HVS)的特性引入到值域块的四叉树划分中,在定义域池的生成以及值域块与定义域块的匹配上采用分类的方法,然后利用彩色图像R,G,B三个分量的相关性,将三色分量图按某种方式组合成一个灰度图像,提出一种使用HVS的彩色图像分形编码算法.实验表明,在保证图像解码质量的情况下,该算法比Fisher和JPEG算法能达到更高的压缩比,并且在编码时间上优于Fisher算法.

Fractal color image coding algorithm based on human visual system

By analyzing the traditional quadtree fractal coding algorithm, a Human Visual System (HVS) based fractal color image coding algorithm was presented in this paper. The feature of HVS was used to quadtree partition of range blocks. Classification method was also used to construct domain pool. Then the correlation of R, G, B components of a color image was used to merge three color components into one. The improved scheme above was used to search for the best domain block for each range block of the merged image. Experimental results indicate that the presented method in this paper has a higher compression ratio than Fishers method and JPEG while remaining good reconstructured image quality, and it is faster than Fisher's method.