A Connected System for the Improvement of Hearing Aids Fitting and Automatic Algorithm Selection
* Presenting author
Progress in hearing aid technology, motivated by the growing number of people affected by hearing-loss, has permitted to improve the speech intelligibility and quality of life of hearing aid users. However, these benefits still come at the price of a fitting period that can be tedious, costly for the acousticians, reduce the satisfaction of the end user and, ultimately, hinder the acceptance of hearing aids. Additionally, even a well fitted hearing aid needs to adapt the speech enhancement algorithm it uses to the encountered acoustic condition. Though most modern hearing aids are already equipped with an acoustic scenes classifier used to determine the best algorithm, this classification is limited and the algorithm selection could be improved by using recent advances in the field of computational auditory scenes analysis. For this purpose, researchers require a large database of labeled data that, contrary to most existing corpora, is tailored to hearing aid applications. The system described in this paper uses the connectivity features of modern hearing aids to both shorten the fitting period and improve the automatic selection of algorithms.