基于相关系数的快速分形图像编码算法

Algorithm for Fast Fractal Image Encoding Based on Correlation Coefficients

自Jacquin提出的分形块编码以来,各种改进算法不断出现,不同程度地减少了编码时间。该文基于相关系数的快速分形图像编码算法,从理论上研究了图像块匹配误差度量和图像块相关系数之间的关系,论证了极小化图像块的均方误差等价于极大化图像块的相关系数,提出了基于相关系数的分形图像编码算法,实现了在解码图像不降质的情况下大大地缩短了编码时间。

Since Jacquin proposed the theory of fractal block coding, a number of improved algorithms were emerged which could reduce the image encoding time in some dgree. The paper proposed the algorithm for fast fractal image encoding based on correlation coefficients： To begin with, proved a equation linking the MSE（mean square error） of the imge tiles with the correlation coefficients of the imge tiles and demonstrated that to minimize their MSE was equal to maximize their correlation coefficients, In succession, the algorithm was designed to advance the theory of the Jacquin coding and fasted the speed of the image encoding without degrading the quality of the imge.