Improved PNN algorithm for initial codebook design of LBG
用于LBG初始码书设计的改进PNN算法

In vector quantization (VQ), the initial codebook design is very important for VQ codebook. This paper introduces two improved PNN algorithms for the design of the initial codebook of LBG algorithm in order to promote the performance and quality of final codebook. The improved algorithm uses sorting training vectors and to the sum of each training vector components and merging multiple training vector iteration. The improved PNN algorithm effectively reduces the complexity and convergence time. Experiments show that the algorithm is reasonable and effective. The proposed method can be combined with the LBG algorithm to further improve the quality of codebook.