REFERENCE	DGICYT - PB94 - 0919
TÍTLE	Implementation of a model of speech recognition based on acoustic features
CHAIR	Dr. Eugenio Martínez Celdrán
DURATION	1995-1998
ABSTRACT	The goal of this research work is the making of a model that describes how the listener processes the acoustic speech signal and succeeds in transforming it into a sequence of phonemes. In order to eliminate an excessive dependence on intuition and the incoherences of any theoretical model, we intend to create an automatic speech recognition computer program which simulates the theoretical model. To carry it out, we will make use of the flow chart which arises from the theoretical model and from the invariant cues already dealt with in the previous project. Finally, the program will be evaluated subjecting it to real speech recognition, using the voices of a minimum of thirty male and female subjects. If the results do not coincide with those we expect, we will remake the program until achieving these results, which could affect the theoretical model at the same time.