Objective measures based on neural networks for hearing loss compensation techniques
Apichat Tungthangthum, Janet C. Rutledge · 2002
An objective measures system has been developed to predict the results of subject-based tests for sensorineural hearing loss compensation techniques. Parameters related to the loudness level of the compensated speech signal are extracted from its frequency spectrum. These parameters are then used to train a neural network based phoneme classifier. Good prediction results have been achieved for two hearing impaired subjects.>