The goal of this DARPA program is to compare wavelet techniques with a variety of competing transforms such as the short-time Fourier transforms, the Wigner-Ville transform, constrained total least squares and Prony single-valued decomposition. The comparison domain is the detection and identification of passive underwater transient acoustic signals. This paper focuses on the comparison process. Preliminary classification results are shown with the current feature set, based on the analysis of a database collected in the laboratory.
展开▼