A self-supervised vowel recognition system

摘要

The paper describes an adaptive model for computer recognition of vowel sounds with the first three formants as features. The method uses a single pattern training procedure for self-supervised learning and maximum value of fuzzy membership function is the basis of recognition.The algorithm with selected representative points and a number of guard zones which are ellipsoidal in the three-dimensional feature space around the representative vectors of the classes is taken as a supervisor. An optimum zone for self-supervised learning is found to correspond to half of the class variances beyond which the machine loses its efficiency. A comparison with a nonadaptive recognition system has also been included.