Vector Quantization Data Compression

doi:10.1117/3.1002297.ch4

Book: Optical Satellite Data Compression and Implementation

Author(s): Shen-En Qian

Published: 2013

https://doi.org/10.1117/3.1002297.ch4

Author Affiliations +

Abstract

Vector quantization (VQ) is an efficient coding technique to quantize signal vectors. It has been widely used in signal and image processing, such as pattern recognition and speech and image coding. A VQ compression procedure has two main steps: codebook training (sometimes also referred to as codebook generation) and coding (i.e., codevector matching). In the training step, similar vectors in a training sequence are grouped into clusters, and each cluster is assigned to a single representative vector called a codevector. In the coding step, each input vector is then compressed by replacing it with the nearest codevector referenced by a simple cluster index. The index (or address) of the matched codevector in the codebook is then transmitted to the decoder over a channel and is used by the decoder to retrieve the same codevector from an identical codebook. This is the reconstructed reproduction of the corresponding input vector. Compression is thus obtained by transmitting the index of the codevector rather than the entire codevector itself.