论文部分内容阅读
有DCT补偿的分形编码可以克服纯分形图像压缩的弱点,但难以实现对输出码流的控制.提出自相似特征矢量(FVS, Feature Vector of Self-resemblance)的概念,并用与基准图像相比较的方法以完成压缩参数的最佳选取.实验结果证明,这种方法在给定压缩比时能得到该压缩比下最佳的恢复图像质量,实现对压缩比的控制.
Fractal coding with DCT compensation can overcome the weakness of pure fractal image compression, but it is difficult to control the output stream. The concept of Feature Vector of Self-resemblance (FVS) is proposed and compared with the reference image to get the best compression parameters. Experimental results show that this method can get the best image quality under the compression ratio and achieve the control of the compression ratio at a given compression ratio.