Image Coding Based on Grey Relational Clustering Vector Quantization

摘要

In this paper, a coding scheme for continuous tone images aiming at high compression ratios via grey relational clustering is performed, in which different classes are chosen to achieve a better quality of the decoded image according to grey theory. Spatial correlation is reduced by a pyramid structured- wavelet decomposition, then prediction of insignificant coefficients is performed across subbands. Afterwards, the reduced set of data is organized in vectors in a way by which residual intra and inter band correlation are exploited. Finally, such vectors are coded by grey relational clustering vector quantization (GRCVQ). The reconstructed images produced by the proposed scheme can even be compared with the JPEG2000 at some rates.