Title :
Design of signal-adapted biorthogonal filter banks
Author :
Lu, Wu-Sheng ; Antoniou, Andreas
Author_Institution :
Dept. of Electr. & Comput. Eng., Victoria Univ., BC, Canada
fDate :
1/1/2001 12:00:00 AM
Abstract :
A method for the design of signal-adapted, M-channel biorthogonal filter banks of finite length is presented. The design problem is formulated as a constrained optimization problem and is solved by converting it into an iterative line-search problem through a first-order parameterization of the perfect reconstruction (PR) constraint. It is also shown for the two-channel case that if the analysis and synthesis lowpass filters are of different lengths, a refinement of the algorithm is possible that leads to a solution in a very small neighborhood of a local minimizer, which satisfies the PR constraint precisely
Keywords :
filtering theory; iterative methods; optimisation; search problems; M-channel biorthogonal filter banks; analysis lowpass filter; constrained optimization problem; design method; finite length; first-order parameterization; iterative line-search problem; local minimizer; perfect reconstruction constraint; signal-adapted biorthogonal filter banks; synthesis lowpass filter; Algorithm design and analysis; Channel bank filters; Constraint optimization; Design methodology; Design optimization; Filter bank; Finite impulse response filter; Matrix converters; Signal design; Transfer functions;
Journal_Title :
Circuits and Systems I: Fundamental Theory and Applications, IEEE Transactions on
