Abstract
A fast image coding scheme built on block truncation, constant, and planar block coding is suggested. The model selection is made to give the best reconstruction. Multi-codebook vector quantization (VQ) of model type and gray levels is used. The selection of model codebook and gray level quantization is based on the desired compression ratio.
Originalsprog | Engelsk |
---|---|
Tidsskrift | Pattern Recognition Letters |
Vol/bind | 21 |
Udgave nummer | 13-14 |
Sider (fra-til) | 1141-1148 |
Antal sider | 8 |
ISSN | 0167-8655 |
DOI | |
Status | Udgivet - dec. 2000 |