序列图像编码中的遗传优化块匹配算法

A GENETIC OPTIMIZATION BLOCK MATCHINGALGORITHM FOR IMAGE SEQUENCE CODING

讨论了序列图像编码的运动估值算法。在剖析传统块匹配算法的基础上，针对其非全局最佳的缺陷，将遗传优化算法的思想应用于分层块匹配算法，提出一种遗传优化块匹配算法。算法分析和计算机模拟表明，此算法可较好地解决现有快速搜索算法中存在的非全局最小点问题，提高数据压缩效率，与穷尽搜索算法相比运算复杂程度明显减小。

In this paper discuss an algorithm of the motion estimation in sequenceimage coding. After analysing traditional Block Matching Algorithms (BMA), an improved hierarchical search algorithm, which adopts the idea of generic optimization algorithm and alleviates the local minimum problem, is presented. Theoretical analysisand computer simulation show that a higher compression rate than those of existingfast-search algorithms is obtained and the computation time compared with full searchalgorithm is also reduced.