Engineering Papers⌕ Search

NASA NTRS · 19930010236

Locally adaptive vector quantization: Data compression with feature preservation

Abstract

A study of a locally adaptive vector quantization (LAVQ) algorithm for data compression is presented. This algorithm provides high-speed one-pass compression and is fully adaptable to any data source and does not require a priori knowledge of the source statistics. Therefore, LAVQ is a universal data compression algorithm. The basic algorithm and several modifications to improve performance are discussed. These modifications are nonlinear quantization, coarse quantization of the codebook, and lossless compression of the output. Performance of LAVQ on various images using irreversible (lossy) coding is comparable to that of the Linde-Buzo-Gray algorithm, but LAVQ has a much higher speed; thus this algorithm has potential for real-time video compression. Unlike most other image compression algorithms, LAVQ preserves fine detail in images. LAVQ's performance as a lossless data compression algorithm is comparable to that of Lempel-Ziv-based algorithms, but LAVQ uses far less memory during the coding process.

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Cheung, K. M., Sayano, M.. 1992-08-15. Locally adaptive vector quantization: Data compression with feature preservation. https://engineeringpapers.com/record/nasa%3A19930010236/locally-adaptive-vector-quantization-data-compression-with-feature-preservation

Citations retain the original authors and DOI and include this site’s record URL. If this discovery service contributed to your research, you can also cite Engineering Papers by Indiciae: BibTeX · RIS.