A Genetic Simulated Annealing Kernel Vector Quantization Algorithm for Image Compression

摘要

In this paper, we approached the problem of vector quantization from a different point of view and proposed a new partition-based evolutionary algorithm for designing vector quantizer. This method is referred to as Genetic Simulated Annealing Kernel Vector Quantization (GSAKVQ) algorithm. The simulated annealing method proposed in this paper makes the candidate solution approach to the optimal solution, and new special crossover operator and mutation operator are present for the partition-based code scheme. Experimental results demonstrate that GSAKVQ algorithm outperforms the traditional evolutionary algorithms in the fields of image compression.