TWO-DIMENSIONAL VECTOR QUANTIZER WITH VARIABLE LENGTH LOSSLESS CODE FOR LAPLACIAN SOURCE

Authors

  • Milan R. Dincic University of Nis
  • Zoran H. Peric University of Nis

DOI:

https://doi.org/10.5755/j01.itc.40.2.428

Keywords:

vector quantizer, lossless code, Golomb-Rice code, execution complexity

Abstract

The main aim of this paper is to apply variable length lossless code on output points of the two-dimensional vector quantizer for Laplacian source. This problem is not yet solved in the literature. In this paper two-dimensional quantizer is designed using Helmert transform and its optimization is done. New lossless code is introduced. It is very simple but very close to the ideal code since it gives the bit-rate very close to the entropy. It is applied on the two-dimensional quantizer. It is shown that our two-dimensional quantizer can achieve the same performances as scalar uniform and nonuniform quantizers, but using much smaller number of points per dimension. Therefore, two-dimensional quantizer has smaller execution complexity.

http://dx.doi.org/10.5755/j01.itc.40.2.428

Downloads

Published

2011-05-31

Issue

Section

Articles