A self-deleting neural network for vector quantization

Abstract

Vector quantization is required the algorithm that minimizes the distortion error, and used for both storage and transmission of speech and image data. For a neural vector quantization, the self-creating neural network and self-deleting neural network and known for showing fine characters. In this paper, we improve the self-deleting neural network, and propose a generalization algorithm combining the creating and deleting neural networks. We discuss algorithms with neighborhood relations compared with the proposed one. Experimental results show the effectiveness of the proposed algorithm.