Improved hierarchical vector quantization for image compression

Abstract

In this paper, we present several contributions on hierarchical vector quantization for image compression: the hierarchical multirate vector quantization, the hierarchical finite state vector quantization, the edge modified vector quantization and the trimming vector quantization. These vector quantization techniques can reach compression ratios as high as 30-70 for monochrome image coding and 60-120 for colour image coding, while the reconstructed images still have satisfactory perceptual qualities. >