A novel entropy-constrained competitive learning algorithm for vector quantization

Wen-Jyi Hwang, Bo Yuan Ye, Shi Chiang Liao

Research output: Contribution to journalArticle

9 Citations (Scopus)

Abstract

In this paper, we present a novel competitive learning algorithm for the design of a variable-rate vector quantizer (VQ). The algorithm, termed entropy-constrained competitive learning (ECCL) algorithm, can achieve a near-optimal performance subject to the average rate constraint. Simulation results show that, under the same average rate, the ECCL algorithm enjoys a better performance than that of the existing competitive learning algorithms. Moreover, the ECCL algorithm outperforms the entropy-constrained vector quantizer (ECVQ) (Chou et al., IEEE Trans. Acoust. Speech Signal Process. 37 (1989) 31-42) design algorithm under the same rate constraint and initial codewords. The ECCL algorithm is also more insensitive to the selection of initial codewords as compared with the ECVQ algorithm. Therefore, the ECCL algorithm can be an effective alternative to the existing variable-rate VQ design algorithms for the applications of signal compression.

Original languageEnglish
Pages (from-to)133-147
Number of pages15
JournalNeurocomputing
Volume25
Issue number1-3
DOIs
Publication statusPublished - 1999 Apr 1

Keywords

  • Competitive learning
  • Signal compression
  • Vector quantization

ASJC Scopus subject areas

  • Artificial Intelligence
  • Cellular and Molecular Neuroscience

Fingerprint Dive into the research topics of 'A novel entropy-constrained competitive learning algorithm for vector quantization'. Together they form a unique fingerprint.

  • Cite this