论文部分内容阅读
本文提出了一种新的基于遗传算法的快速运动估计方法 .该方法对遗传算法进行了改进 ,采用“阈值法”确定选择算子 ,并将基因变异所导致的随机搜索与特定目标搜索相结合 ,解决了以往快速搜索算法易陷于局部最优的问题 ,大大提高了运动估计速度 .该方法还将运动矢量空间一致性原则用于初始种群的选取 ,进一步提高了算法性能 .由于其具备遗传算法固有的规则性和高度并行性 ,该方法适合于采用VLSI实现实时视频编码器 .
In this paper, a new fast motion estimation method based on genetic algorithm is proposed, which improves the genetic algorithm, uses the “threshold method” to determine the selection operator, and combines the random search caused by the genetic mutation with the specific target search , Which solves the problem that the fast search algorithm is easily trapped in the local optimum in the past and greatly improves the speed of the motion estimation.This method also uses the principle of spatial consistency of motion vector for the initial population selection to further improve the performance of the algorithm.Because of its genetic algorithm Inherent regularity and high degree of parallelism, this approach is suitable for implementing real-time video encoders using VLSI.