一种基于压缩感知的快速混合分形编码算法研究

A compress sensing based fast Hybrid fractal coding algorithm

针对分形编码的方块效应以及编码时间过长的问题,本文提出分形编码与压缩感知的混合算法。该算法只对低频子图进行分形编码,减少编码时间;再对低频差值图像以及其他子图进行压缩感知编码,以补充分形解码后图像所包含的信息。实验表明:对全局图像的小波变换缓解方块效应,只对低频子图进行分形编码大大提高编码速度,对差值图像以及其他子图的再编码提高了整个重构图像的质量。

To overcome the shortcoming of long coding time and blocking effect of the fractal compress coding algorithm, a fractal-CS （ compress sensing ） hybrid coder is proposed in this paper. After a Dabechies9/7 transformation to the original image, fractal compression algorithm is used to coding the low-frequency sub-image, then a subtraction is made between decoder low-frequency sub-image and original low-frequency sub-image, and coding the combined image by CS coding algorithm. Finally, decode the combined image, fuse the decoded combined image and decoded low-frequency sub-image to get the reconstructed image. The experimental results show that the proposed algorithm not only reduces the blocking effect and improve code speed, but also obtains high quality reconstructed image.