Transform coding of monochrome and color images using trellis coded quantization

Abstract

Trellis coded quantization (TCQ) is incorporated into a transform coding structure for encoding monochrome and color images at 1 bit/pixel. Both fixed-rate and entropy-constrained designs are considered. For monochrome images, a simple 4-state fixed-rate TCQ-based system provides gains in peaksignal-to-noise ratio of up to 2.4 dB over scalar quantizer based systems, whale the entropy-constrained design provides gains of up to 7.0 dB. Gains in PSNR for color images of up to 1.6 dB and 3.9 dB are achieved for the 4-state fixed-rate and entropyconstrained TCQ-based systems, respectively. The subjective quality of the images produced by the TCQbased systems is significantly better than that of the scalar quantizer-based systems. The high frequency background noise, fuzziness, and block artifacts produced b y the scalar quantizer-based systems are virtually eliminated by the TCQ-based systems.