Design of lossless block transforms and filter banks for image coding

Lossless block transforms and filter banks that map integers to integers ar e important for unified lossless and lossy image coding systems. In this pa per, we present simple yet effective methods for designing lossless version s of block transforms and FIR filter banks. First, an N-point lossless tran sform and a lossless interpolative prediction are introduced. Next, we demo nstrate that filter banks can be decomposed into 2-point transforms or inte rpolative predictions. Lastly, lossless versions of block transforms and fi lter banks are obtained by replacing every constituent module by the corres ponding lossless version. Lossless versions of 8-point discrete cosine tran sform (DCT), 8-point Walsh-Hadamard transform (WHT) and several filter bank s are designed and their lossless compression performance is evaluated.