基于改进粒子群算法的矢量量化码书设计研究

提出一种粒子群分组并行寻优码书设计算法,应用于图像的矢量量化编码中.可以得到性能较好的码书.利用同一种群两个分组分别进化,同时相互监督,某一个分组或者两个分组都陷入局部最优时,它能够通过相互作用跳出局部最优;然后通过对训练矢量进行排序,合理选择初始码书,使码字的分布更加合理,增强搜索多样性;最后通过仿真实验验证了该改进算法的合理性.

Codebook Design of Vector Quantization Based on Improved Particle Swarm Optimization

This paper presents swarm optimization algorithm based on a pair of parallel particles, which can be used to get a good codebook in the vector quantization of image coding. Using the two groups evolve separately and monitoring each other, when one or two groups fell into local optimum, it could jump out of local optimum through interaction; By sorting the training vector, we selected the initial codebook to make the code distribution more reasonable and enhance the diversity of search; Finally, the simulation results prove that the improved method is reasonable.