CHOOSING OPTIMAL ORTHOGONAL WAVELETS FOR SIGNAL APPROXIMATION
用于信号逼近的最佳正交子波选择方法

The discrete wavelet transform decomposes a discrete time signal into an approximation sequence and a detail sequence at each level of resolution. Compactly supported orthonormal wavelets correspond to perfect reconstruction (PR) quadrature mirror filter (QMF) banks. This paper deals with the problem of choosing orthogonal wavelet (scaling) filters for best signal approximation at some scales. By using a kind of parametrization method, the constrained optimization can be converted into an unconstrained one. Some simulations are shown here.