Vector quantization codebook designing method based on genetic algorithm

Abstract

The invention relates to a vector quantization codebook designing method based on a genetic algorithm, which comprises the following steps coding chromosomes; generating an initial population; calculating the adaptability of each chromosome; selecting an individual which enters a next generation; making the chromosomes intersect; making the chromosomes vary; and generating a new population. The vector quantization codebook designing method aims at the population. A selection operator is used for performing targeted optimum operation on the populations. Furthermore diversity of the populations is increased through an improved crossover algorithm. Finally the average adaptive value of the population is improved through a mutation operator, so that the average adaptive value gets rid of a partial least point. The vector quantization codebook designing method has a good partial searching capability through reserving an LBG optimization selection strategy.

Description

technical field [0001] The invention relates to the technical field of image compression coding, in particular to a method for designing a vector quantization codebook based on a genetic algorithm. Background technique [0002] Vector quantization technology includes codebook design, codeword search and codeword index matching, among which codebook design is the key to determine the compression effect. Linde, Buzo and Gray proposed K-means clustering algorithm for codebook design in 1980, which is also called LBG algorithm. The LBG algorithm is rigorous in theory, simple and convenient to implement, and has been successfully applied to image vector quantization and compression. The LBG algorithm is an optimal codebook iterative algorithm based on the following two necessary conditions, namely the nearest neighbor condition and the center condition. The nearest neighbor condition ensures the shortest Euclidean distance between the codeword and each vector; the center condit...

AI Technical Summary

Problems solved by technology

However, the classic LBG algorithm has fatal shortcomings:

[0003] (1) The requirements for the initial codebook are high, and the convergence speed and final codebook performance are sensitive to the initial codebook;

[0004] (2) It cannot stably converge to the global optimal solution, and generally falls into a local optimal solution;

[0005] (3) The amount of calculation is huge, the time and space calculation complexity is high, and the codebook generation speed is slow;

[0006] (4) The generated codebook is out of order, which leads to the complexity of the subsequent codebook search algorithm

Images