基于相关系数的快速分形图像编码

分形图像编码是一种有效压缩技术。在分形编码中，一幅图像由一个不动点接近原始图像的压缩变换来编码，然后由源于Banach不动点定理的迭代过程来解码。该文提出了快速分形图像编码的一种基于相关系数的编码方案，不需要改变现有的分形解码过程。该方案基于这样的假设，两个等尺寸的子块不能组成匹配对，除非它们的相关系数相对较大。它能够以直接的方式融入其它的分形编码算法。计算机仿真显示，对8幅复杂性不同的测试图像，该文算法能够平均加快编码3倍或更多，同时PSNR(peak signal-to-noise ratio)平均下降不到0．1dB，且主观质量有时甚至好于基本分形算法。

Fast Fractal Image Encoding Based on Correlation Coefficients

Fractal image coding is an efficient compression technique, in which an image is encoded by a contractive transformation whose fixed point is close to the original image, and then is decoded by using the iteration procedure stemmed from the well-known Banach fixed-point theorem. This paper proposes a correlation-coefficients-based scheme for fast fractal image encoding, which does not need to change the existing fractal decoding procedure. The proposed scheme is based on the hypothesis that two equal-sized image blocks cannot be closely matched unless their correlation coefficient is relatively large. It can be employed by the other fractal image encoding algorithms in a straightforward manner. Computer simulations on 8 test images with different complexities demonstrate that the proposed scheme could achieve averagely a speedup of 3 times or more, while it averagely gives an insignificant decrease of less than 0.10dB in the PSNR (the peak signal-to-noise ratio) and sometimes subjective qualities of the decoded image are even better than that using the corresponding baseline fractal algorithm.