Feature extraction from wavelet coefficients for pattern recognition tasks

In this paper, a new efficient feature extraction method based on the fast wavelet transform is presented. This paper especially deals with the assess ment of process parameters or states in a given application using the featu res extracted from the wavelet coefficients of measured process signals. Si nce the parameter assessment using all wavelet coefficients will often turn out to be tedious or leads to inaccurate results, a preprocessing routine that computes robust features correlated to the process parameters of inter est is highly desirable. The method presented divides the matrix of compute d wavelet coefficients into clusters equal to rowvectors. The rows that rep resent important frequency ranges (for signal interpretation) have a larger number of clusters than the rows that represent less important frequency r anges. The features of a process signal are eventually calculated by the eu clidean norms of the clusters. The effectiveness of this new method has bee n verified on a flank wear estimation problem in turning processes and on a problem of recognizing different kinds of lung sounds for diagnosis of pul monary diseases.