ICA, a key to solve the puzzle of hearing $(Aˋ

[Guoping Li <lgp@xxxxxxxxxxxxxxxx>]

Dear all,

I came across the term 'ICA'(Independent component analysis), which could be used to separate multiple independent sources (Poerre Comon,1994; Paris Smaragdis,2001).  ICA now is used very widely in biomedical signal processing (EEG, EMG...) as well as image processing. Paris $(Aㄗ2001ㄘshowed that ICA could be used as an unify platform for the perceptual framework of hearing (preprocessing, grouping and scene analysis).

The deaf, even with the  hearing aid or cochlear implant,  has difficulties of understanding and separating the sound sources in adverse environment, such as in a 'cocktail party'.  And It seems the present  signal processing technique  used in hearing aid or cochlear implant provides  little help in such situation. 


 ICA was designed to solve the problem of cocktail party effect using  higher order statistical methods. Is it possible that we can use this technique and implement it in the hearing aid or cochlear implant? Is there any paper which is relared to the implementation of ICA in the hearing aid or cochlear implantˋ

I would be very grateful if anybody could  make some comments or provide related resources.


Thanks.

Guoping LI
Institute of Sound and Vibration Research
Hearing and Balance Center
University of Southampton, SO17 1BJ, UK
Tel 023 80592842