Design of linear-phase two-channel quadrature mirror filter banks using neural minor component analysis

This paper presents a neural-based minor component analysis algorithm for designing quadrature mirror filter banks in the least-squares sense. The objective function of the design problem is formulated as an eigenvector solving problem of a real, symmetric and positive-definite matrix. An alternative minor component analysis algorithm based on neural learning rule is exploited to achieve the eigenfilter design of the quadrature mirror filter bank with accurate performance.