Application of principal component filter banks to subband coding

Signal-adaptive multi-rate filter banks have received a growing interest in the research community. The filter banks are designed to match the statistical properties of the input signals. In particular, a principal component filter bank (PCFB) attains maximal energy compaction in the sense that the mean squared error between the input and output (when only a limited number of subbands are retained for reconstruction) is minimized, hence maximizing the theoretical coding gain (TCG). While the theoretical aspects have been developed in detail previously, this paper discusses the application of PCFBs to subband coding of images.