Image sequence coding using adaptive nonuniform tree-structured vector quantization
Journal
Journal of Visual Communication and Image Representation
Journal Volume
2
Journal Issue
2
Pages
166--176
Date Issued
1991-06
Author(s)
Abstract
Since nonuniform tree-structured vector quantization (TSVQ) has been shown to have better performance than uniform TSVQ in coding speech, we apply the nonuniform TSVQ to encoding images. In the encoding image sequence for transmission or storage, the coding algorithm must have the ability to adapt to changing image characteristics from sequence to sequence and from frame to frame. This paper describes a new adaptive nonuniform TSVQ in which the codebook tree is reorganized at an update interval by using the proposed splitting and shrinking operations. On the CCITT test sequence "CLAIRE", gains of up to 2.1 dB in SNR are realized for adaptive nonuniform TSVQ over static TSVQ, resulting in high image quality at 1.115 bits per pixel. © 1991.
Type
journal article