Wavelet packet based transient signal classification

Nonstationary signals are not well suited for detection and classification by traditional Fourier methods. An alternate means of analysis needs to be used so that valuable time-frequency information is not lost. The wavelet packet transform is one such time-frequency analysis tool. The authors summarize efforts which examine the feasibility of applying the wavelet packet transform to automatic transient signal classification through the development of a classification algorithm for biologically generated underwater acoustic signals in ocean noise. The formulation of a wavelet-packet based feature set specific to the classification of snapping shrimp and whale clicks is given