A Self-Deleting Neural Network for Vector
Michiharu Maeda · 1996
Vector quantization is required the algorithm that minimizes the dist#ortion error, and used for both storage and transmission of speech and image data. For a neural vector quantization (l), the self-creating neural network (2) and self-deleting neu- ral network (3) and known for showing fine charac- ters. In this paper, we improve self-deleting neural network, and propose a general ion algorithm com- bining the creating and deleting neural networks. We discuss algorithms with neighborhood relations (2)~(5) compared with the proposed one. Experimental re- sults show the effectiveness of the proposed algorithm.