COMPARISON OF DIFFERENT METHODS OF CLASSIFICATION IN SUBBAND CODING OF IMAGES

This paper investigates various classification techniques, applied to subband coding of images, as a way of exploiting the nonstationary nat ure of image subbands, The advantages of subband classification are ch aracterized in a rate-distortion framework in terms of ''classificatio n gain'' and overall ''subband classification gain.'' Two algorithms, maximum classification gain and equal mean-normalized standard deviati on classification, which allow unequal number of blocks in each class, are presented, The dependence between the classification maps from di fferent subbands is exploited either directly while encoding the class ification maps or indirectly by constraining the classification maps, The trade-off between the classification gain and the amount of side i nformation is explored, Coding results for a subband image coder based on classification are presented, The simulation results demonstrate t he value of classification in subband coding.