Audio signal identification via pattern capture and template matching

This research reports on a system able to classify different signals contai ning auditive information based on capture of small signal segments present in specific types of sound. After using a Haar wavelet transform at the pr eprocessing stage, a neural network known as the O-algorithm compares segme nts from candidate audio signals against predefined templates stored in the network. The classification performance is tested with three different app lications. (C) 2000 Elsevier Science B.V. All rights reserved.