Non-Parametric Vector Quantization Algorithm

Main Article Content

Haemwaan Sivaraks
Athasit Surarerks

Abstract

Recently researches in vector quantization domain concentrate on improving the performance of parametric algorithms (i.e., need to specify either the codebook size or the expected distortion.) The specification of the both parameters gives the users some difficulties. We, in the research, propose a nonparametric vector quantization (NVQ) algorithm. The concept is to manage Voronoi regions with respect to the distortion decreasing rate. Experimental results point that the average distortions are statistically decreased comparing with LBG and adaptive incremental LBG algorithms at the 99% of confidence level. Both speech and image data are also included
in our consideration.

Article Details

How to Cite
[1]
H. Sivaraks and A. Surarerks, “Non-Parametric Vector Quantization Algorithm”, ECTI-CIT Transactions, vol. 3, no. 1, pp. 31–38, Apr. 2016.
Section
Artificial Intelligence and Machine Learning (AI)