Filter bank design for redundant wavelet transforms

This paper introduces a technique for the design of a general class of filter bank for compactly supported wavelets. The technique allows the design of real or complex wavelet transforms that have linear phase and generate redundant expansions. The design is based on the factorization of paraunitary matrices and a nonlinear unconstrained optimization formulation. In the particular case of two-channel critically sampled filter banks, the design can generate the Daubechies' wavelets filter banks.