A Neural Ensemble For Speech Recognition
Mario Costa, Paolo Gay, J.D. Palmisano, Eros Gian Alessandro Pasero · 2005
Automatic speech recognition is a severe application for neural networks. An ensemble of more neural networks can be the keystone to increase the performance of the recognizer. Different techniques to pre-process the vocal signal are also shown. This approach is used to implement a “viva voice” recognizer for a car phone. A PC based demonstrator is available to recognize digits independently from the speaker.