Vector Quantization Codebook Generation Method Using Classification Performance

摘要

The present invention relates to a method of generating a codebook used for vector quantization based on classification performance of a center vector of a group to be classified. In particular, the method selects the largest average distortion among the codewords in the codebook. A method of generating a vector quantization codebook that reduces overall distortion by repeatedly performing a vector quantization process between a training vector and currently generated codewords when the number of codewords in a codebook reaches a target value. Is disclosed.