Planar-oriented ripple based greedy search algorithm for vector quantization

Abstract

Vector quantization techniques have been used in various applications. The efficiency of search algorithm is very important for vector quantization. In this paper, a planar-oriented ripple based greedy search algorithm is proposed to reduce the search time of vector quantization. In order to reduce the dimensions of vectors, principal component analysis is used to find the principal components with the most variability. To find the closer codewords, Voronoi diagram is applied to find the Voronoi cells, then, the adjacency list can be generated. Finally, to improve the efficiency of codeword searching, a greedy search is adopted to reduce the searching space. The results of the present study show that the proposed approach achieves a better performance than planar Voronoi diagram search algorithm.