Vector code quantizer based on particle group

Abstract

The invention relates to an efficient data compression technique, in particular to a vector code book quantizer which is used in speech coding and image compressing systems and is based on particle swarm. The technique of the invention is characterized in that the vector code book quantizer is designed according to the combined proposal of particle swarm optimization and simulated annealing; in the process of designing the vector code book quantizer, an empty cell processing proposal newly posted is adopted to solve the problem of empty cells. The vector code book quantizer based on the particle swarm of the invention can be realized by software simulation. According to the evaluation of unofficial hearing tests of speech which is obtained after the vector code book quantizer being coded and reconstructed through the adoption of speech data, the speech reconstructed by the newly posted method is better than the speech reconstructed based on the particle swarm optimization no matter in the aspect of clarity, naturalness or comprehensibility.

Description

Technical field: [0001] The invention relates to a high-efficiency data compression technology, in particular to a particle swarm-based vector codebook quantizer used in speech coding and image compression systems. Background technique: [0002] As we all know, in vector quantization technology, codebook quantizer design is one of the key technologies. An optimal codebook quantizer design algorithm—LBG algorithm proposed by Linde, Buzo and Gray in 1980. Because of its theoretical rigor, relatively fast convergence speed, simple implementation process and good practical results, it is widely used in speech and image processing applications. However, the algorithm still has the disadvantages of large amount of calculation, easy to fall into local optimal solution, and its performance strongly depends on the selection of the initial codebook. In response to these problems, scholars began to propose various improved algorithms, such as simulated annealing codebook design algor...

AI Technical Summary

Problems solved by technology

[0009] ① In the case that the sub-neighborhood is divided into empty, the sub-neighborhood method will randomly select a training vector to replace the codeword of the empty cell cavity as a new codeword, which lacks a certain theoretical basis and will lead to the corresponding codeword word uncertainty

[0010] ② If only one training vector is clustered in a certain cell, and this training vector happens to be classified into the sub-neighborhood division of an empty cell cavity, then follow the strategy 2 in the sub-neighborhood method to When processing this empty cell, if the training vector is classified into this empty cell, the original cell of the training vector will become an empty cell, which will increase the distortion and reduce the effect of clustering

[0011] ③ Blindly classify all secondary codewords into the next adjacent cavity, which will increase the overall distortion to a large extent and affect the performance of the final codebook quantizer

Images