改进转动惯量特征的快速分形图像编码算法

全搜索分形图像编码过程特别耗时的原因在于，每个range块都需要在一个很大的domain块池里寻找最佳匹配domain块。为了改进这个缺点，重新定义了图像规范块的转动惯量特征，证明了它与匹配均方根误差间的关系不等式，据此提出了一个限制搜索范围来加快编码过程的算法：一个待编码range块的最佳匹配块搜索范围仅在与它的转动惯量特征值相近的domain块的邻域内搜索，邻域半径的大小由预先设置的误差阈值来确定。三幅图像的仿真结果表明，它确实能够在不降低解码图像质量的情况下，通过减少搜索范围达到了平均加快全搜索分形编码算法的编码速度26倍左右（误差阈值为10），且也优于转动惯量算法和三均值特征算法。

Fast fractal image encoding algorithm based on improved moment of inertia feature

Fractal image encoding with full search typically requires a very long runtime, which is essentially spent on searching for the bestmatched block to an input range block in a large domain pool. This paper thus proposes an effective method to im prove the drawback, which is mainly based on certified inequality linking the rootmeansquare and Improved Moment of Inertia （IMI） feature of normalized block. During the search process, the IMI feature is utilized to confine efficiently the search space to the vicinity of the domain block having the closest IMI feature to the input range block being encoded, aiming at reducing the searching scope of similarity matching to accelerate the encoding process. Besides, a beforehand error threshold is used to deter mine the size of search neighbourhood. Simulation results of three standard test images show that the proposed scheme not only reduces the searching scope of bestmatched to averagely obtain the speedup of 26 times or so by error threshold set 10, but also can obtain the same quality of the decoded images as the baseline algorithm with full search. Moreover, it is better than the moment of inertia algorithm and the threemean feature algorithm.