基于遗传算法的图像序列译码性能分析

Fano算法在码树上始终只跟踪一条路径,因此遭遇到比较严重的噪声干扰时,容易偏离正确路径,并且难以在较短时间内回到正确路径上来.本文提出了一种基于遗传算法的纠错序列译码算法,将遗传算法具有的全局优化的迭代进化用于序列译码中,使得译码序列不仅保留最大似然路径,也扩展一些似然度比较小的路径,使被选中的路径具备多样性,可以搜索比较宽的范围,从而尽可能跟踪到最佳路径.同时,能根据干扰的大小通过门限调节自适应地改变群体规模,达到正确而且快速搜索译码的目的.全文讨论了该算法的基本原理和实现可行性,并将这种算法用于视频图像纠错,给出了实验结果及其性能分析.

The Performance Evaluation of the Image Sequential Decoding Based on Genetic Algorithm

Fano Algorithm only searches one path,and it may easily lose correct path when encountering severe noise interference.A new sequential decoding method based on Genetic algorithm is proposed.This method uses globally optimized iterative evolution search strategy and processes several paths simultaneously,so as to not only keep the maximum likelihood path but also extend some less likely paths. As a result,it can search wider range and trace the path optimally;it also can adjust the threshold self-adaptably to change the population size in order to decode correctly and rapidly.The basic theory and feasibility of its realization are discussed and used in the error correction of the video data transmission.The experimental results and performance evaluation have also been given.