Vector Quantisation

Article type: Research Article

Authors: Rak, Remigiusz J.

Affiliations: Electrical Faculty, Warsaw University of Technology, Warsaw, Poland. email: rakrem@iem.pw.edu.pl

Note: [] Address for correspondence: Electrical Faculty, Warsaw University of Technology, PI. Politechniki 1, 00-665 Warsaw, Poland

Abstract: In the last few years Vector Quantisation has become an important technique in speech and image compression and recognition. The jump from one dimension to the multiple ones allows wealth of new ideas, concepts, techniques and applications. In this paper there is enclosed a desription of vector quantisation idea, the design techniques, optimization methods and practical implementations.

Keywords: vector quantisation, codebook, segmentation, transform coding, stochastic relaxation, simulated annealing, training sequence

DOI: 10.3233/FI-1998-34404

Journal: Fundamenta Informaticae, vol. 34, no. 4, pp. 397-412, 1998