NN-based automatic sound classifier for digital hearing aids

This paper centers on exploring proper training algorithms for multilayer perceptrons (MLPs) to be used within digital hearing aids. One argument usually considered against the feasibility of neural networks on hearing aids consists in both their computational complexity and the hardware constraints the hearing aids suffer from. Within this framework, this work focuses on studying the influence of a number of training methods for an MLP able to automatic classify the sounds entering the hearing aid into three classes: speech, noise and music. The training methods explored are Gradient Descent, Levenberg-Marquardt, and Levenberg-Marquardt with Bayesian Regularization. Our results show how the proper selection of the training algorithm leads to a good mean probability of correct classification of 91.7% along with a low number of neurons, the computational complexity being thus reduced. These results have been successfully compared to those obtained from a k-Nearest Neighbors algorithm, which exhibits poorer performance.

[1]  Enrique Alexandre,et al.  Automatic Sound Classification for Improving Speech Intelligibility in Hearing Aids Using a Layered Structure , 2006, IDEAL.

[2]  Paul Mermelstein,et al.  Experiments in syllable-based recognition of continuous speech , 1980, ICASSP.

[3]  R. Gil-Pita,et al.  A hearing aid simulator to test adaptive signal processing algorithms , 2007, 2007 IEEE International Symposium on Intelligent Signal Processing.

[4]  Malcolm Slaney,et al.  Construction and evaluation of a robust multifeature speech/music discriminator , 1997, 1997 IEEE International Conference on Acoustics, Speech, and Signal Processing.

[5]  Xavier Serra,et al.  Towards Instrument Segmentation for Music Content Description: a Critical Review of Instrument Classification Techniques , 2000, ISMIR.

[6]  Mohammad Bagher Menhaj,et al.  Training feedforward networks with the Marquardt algorithm , 1994, IEEE Trans. Neural Networks.

[7]  François Pachet,et al.  Representing Musical Genre: A State of the Art , 2003 .

[8]  G Keidser.  The Relationship Between Listening Conditions and Alternative Amplification Schemes for Multiple Memory Hearing Aids , 1995, Ear and hearing.

[9]  David G. Stork,et al.  Pattern Classification , 1973 .

[10]  Enrique Alexandre,et al.  Exploring the feasibility of a two-layer NN-based sound classifier for hearing aids , 2007, 2007 15th European Signal Processing Conference.

[11]  Martin T. Hagan,et al.  Gauss-Newton approximation to Bayesian learning , 1997, Proceedings of International Conference on Neural Networks (ICNN'97).

[12]  Enric Guaus,et al.  A Non-linear Rhythm-Based Style Classifciation for Broadcast Speech-Music Discrimination , 2004 .

[13]  Mohamed S. Kamel,et al.  On the optimal number of hidden nodes in a neural network , 1998, Conference Proceedings. IEEE Canadian Conference on Electrical and Computer Engineering (Cat. No.98TH8341).

[14]  George Tzanetakis,et al.  MARSYAS: a framework for audio analysis , 1999, Organised Sound.

[15]  G Keidser.  Selecting different amplification for different listening conditions. , 1996, Journal of the American Academy of Audiology.

[16]  Christopher M. Bishop,et al.  Neural networks for pattern recognition , 1995 .

[17]  P. Balaban,et al.  A Modified Monte-Carlo Simulation Technique for the Evaluation of Error Rate in Digital Communication Systems , 1980, IEEE Trans. Commun..

[18]  Stan Davis,et al.  Comparison of Parametric Representations for Monosyllabic Word Recognition in Continuously Spoken Se , 1980 .