Steepest Descent Algorithm For Block Wavelet Transform Filter Coefficients That Generate Orthogonal Transform Matrix

Abstract

It's known that the transform matrix generated by Block Wavelet Transform (BWT) is orthogonal. Eigenvector matrix of the autocorrelation matrix obtained in Karhunen-Loeve transform is also orthogonal. This matrix is used for computing the features of signals. It's thought that BWT can be used in the same manner. A BWT obtained by finite number of filter coefficients will be advantegous. In this paper, a method for the computation of BWT filter coefficients by Steepest Descent algorithm will be proposed.