BibTeX
@MISC{Parks90vectorquantization,
author = {Thomas M. Parks},
title = {Vector Quantization Code Book Design Using Neural Networks},
year = {1990}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
The Kohonen Self-Organizing Feature Map algorithm is compared to the K-Means vector quantization algorithm. Computation and storage requirements are calculated for both algorithms. A new algorithm which takes advantage of the structured code book produced by Kohonen's algorithm is introduced. This algorithm offers a significant computational savings over full-search vector quantization without imposing a storage cost penalty. The results of simulation studies are presented and the performance of the algorithms is compared. 1 Introduction This paper evaluates the use of neural network algorithms for vector quantization. Section 2 gives a brief overview of vector quantization, followed by descriptions of the K-Means algorithm in Section 3 and the Kohonen SelfOrganizing Feature Map in Section 4. Computation and storage costs of these two algorithms are compared in Section 5. Simulation results demonstrating the performance of the algorithms are presented in Section 6. Sec- This repor...
Keyphrases
vector quantization k-means vector quantization algorithm neural network algorithm storage cost structured code book simulation result kohonen selforganizing feature map significant computational saving simulation study full-search vector quantization new algorithm storage cost penalty brief overview k-means algorithm storage requirement kohonen self-organizing feature map algorithm
