中文摘要：

基本分形图像压缩编码算法虽然是一种很有前途的限失真编码方法，但是它存在着编码时间较长、计算复杂度较高的缺点。为了解决分形图像压缩编码算法编码时间过长的问题，基于图像的父块特征，提出了一种改进算法。该算法利用图像父块的几何特征预先把父块库分成Ds、De、Dm三大类，通过在各个类中运用相应的特征将搜索范围限制在与子块特征值相近的邻域内，即将类内全局搜索最佳匹配块转化为类内局部搜索最佳匹配块，有效地减少了搜索对象，从而进一步加快了编码速度。应用该算法与其他算法进行了多次仿真对比实验。实验结果表明，相对于其他算法，在保证一定重建图像质量的前提下，所提出算法的图像编码时间明显缩短，较为显著地提高了算法编码的速度。

英文摘要：

The basic fractal image compression encoding algorithm is a finite lossless encoding method that has great significance, but it spends more time on encoding and is more complicated to calculate. In order to reduce the fractal image compression encoding time,an improved algorithm based on the characteristics of the parent block has been proposed in which the parent block is divided into three cate- gories with the image block geometric features in advance, Ds , De , and Dm . Though use of corresponding features in each class, the search range is limited to the neighborhood closed to the sub-block characteristic values, which means that the best matching block of global search in class is turned into local search for the best matching block. The proposed algorithm has effectively reduced the search ob- jects, which can further accelerate the speed of coding. Test simulations for multiple comparisons have been conducted with the proposed algorithm and others. Simulation results show that compared with other ones,the image encoding time of the proposed algorithm is signifi- cantly shortened in the guarantee of the quality of the reconstructed image, which has more significantly improved the encoding speed.