ABSTRACT
The initial codebook selection of the LBG algorithm will affect the performance of the final codebook, and the design of the vector quantization codebook with basic particle swarm optimization algorithm is prone to local optimal solution problems. Vector quantization includes three major techniques: codebook design, code word search, and code word index allocation. In the 1990s, two scholars in the United States proposed a Particle Swarm Optimization (PSO) algorithm on the basis of the observation of food behavior in birds. Once the inertia weight is added to a particle swarm algorithm, it is called the standard particle swarm optimization algorithm. Although the PSO algorithm uses real numbers to encode vectors, however, because the objects manipulated in vector quantization are all vectors, the codebook design based on PSO is based on the clustering center coding method. The searching and clustering are continued until the codebook with better performance is generated.
