Multilayer feedforward networks (MFN) are tested in detection and classification tasks. The classification of the speech signal into voiced, unvoiced, and silence (V/U/S) provides a preliminary acoustic segmentation of speech, which is important for speech analysis and may also have applications in speech-data mixed communication systems. The classification was made using a MFN with different hybrid features. The paper evaluates the efficiency of these feature vectors, the classification rate as a function of signal to noise ratio and simulation results are presented.
F. Albu, A. Mateescu, "Application of Multilayer Feedforward Network to the Voiced-Unvoiced-Silence Detection Problem", Proceedings of International Symposium on Communications'96, Bucharest, Romania, pp. 532-537, November 1996