Loss of Perfect Reconstruction in Multidimensional Filterbanks and Wavelets Designed via Ex.pdfVIP

IEEE SIGNAL PROCESSING LETTERS, VOL. 4, NO. 10, OCTOBER 1997 295 Loss of Perfect Reconstruction in Multidimensional Filterbanks and Wavelets Designed via Extended McClellan Transformations Dong Wei, Student Member, IEEE, Brian L. Evans, Senior Member, IEEE, and Alan C. Bovik, Fellow, IEEE Abstract— Two extensions of the McClellan transformation transformation, the two extended McClellan transformations have been proposed to transform even-length symmetric one- suffer from a limitation in the design of multidimensional two- dimensional (1-D) prototype filters into even/odd-length multi- channel FIR PRFB’s and wavelets in the sense that there does dimensional filters. We show that, unlike the original McClel- lan transformation, the two extended McClellan transformations not exist any transformation function that can preserve PR suffer from a limitation in the design of multidimensional two- after transformation. For one of the transformations, we point channel finite impulse response (FIR) perfect reconstruction (PR) out that near-PR filter banks are possible. filterbanks and wavelets in the sense that there does not exist any transformation function that can preserve PR after transforma- tion. However, for one of the transformations, we point out that II. EXTENDED MCCLELLAN near-PR filterbanks are possible. TRANSFORMATIONS IN THE DESIGN OF PRFB’S In this letter, boldfaced letters denote -D vectors, e.g., I. INTRODUCTION A