Subject: Re: ICA, From: Paris Smaragdis <paris(at)MEDIA.MIT.EDU> Date: Fri, 25 Mar 2005 13:24:05 -0500If you google for independent component analysis and hearing aids you'll get plenty of relevant references. There have been a lot of papers on this subject and people are actively working on it. Most work I've seen has been using multichannel formulations, which given the prior work on beamforming for hearing aids seemed like a natural thing. Single channel formulations are a bit more complex and will probably take their time to debut on hearing aids. Integrating any of the other more "perceptual" things you can do with ICA on a hearing aid requires a lot more work on interfacing with the human hearing system. Regardless, using any such separation algorithm on a hearing aid is an ill-defined concept. What is needed is a source separation algorithm that will separate what you want to listen to. And quite frankly I'd be very impressed if anyone can come up with an algorithm that can predict a user's auditory focus. Paris Guoping Li wrote: > 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 > >