Performance investigation on DTCWT based on the common factor technique

This paper investigates the performance on a class of DTCWTs based on the common factor technique. The common factor technique was first proposed in [6] by Selesnick, where two corresponding scaling lowpass filters are required to satisfy the half-sample delay condition, and then was modified to improve the analyticity [9] and frequency selectivity [8]. The analyticity of complex wavelet can be improved by minimizing the phase error in the approximation band of allpass filters with given flatness condition. The frequency selectivity of scaling lowpass filters can be improved by locating a specified number of zeros at z = -1 and applying the Remez exchange algorithm to minimize the magnitude response in the stopband. How to determine the approximation band and stopband will influence the performance of DTCWT. Therefore, this paper is dedicated to how to choose the approximation band and stopband properly.