Automatic User Preferences Selection of Smart Hearing Aid Using BioAid

• DOI: 10.3390/s22208031
• Publication Date: 2022-10-21T04:34:30Z
• Authors: Hafeez Ur Rehman ...; Adil Ali Saleem; Muhammad Amjad Raza; Kainat Zafar; Riccardo Russo; Sandra Dudley
• Keywords: Hearing aid; Python

Noisy environments, changes and variations in the volume of speech, non-face-to-face conversations impair user experience with hearing aids. Generally, a aid amplifies sounds so that hearing-impaired person can listen, converse, actively engage daily activities. Presently, there are some sophisticated algorithms available operate on numerous frequency bands to not only amplify but also provide tuning noise filtering minimize background distractions. One those is BioAid assistive system, which an open-source, freely downloadable app twenty-four settings. Critically, this device, suffering loss must manually alter settings/tuning their device when surroundings scene order attain comfortable level hearing. However, manual switching among multiple settings inconvenient cumbersome since forced switch state best matches every time auditory environment changes. The goal study eliminate automate classification algorithm system automatically identifies user-selected preferences based adequate training. aim acoustic recognize audio signature one predefined classes represent it was recorded. BioAid, open-source biological inspired algorithm, used after conversion Python. proposed method consists two main parts: scenes selection experiences. DCASE2017 dataset utilized for classification. Among many classifiers were trained tested, random forests have highest accuracy 99.7%. In second part, clean speech audios from LJ combined scenes, asked listen resulting adjust presets subsets. A CSV file stores subsets at hear clearly against scenes. Various preferences. After training, convolved fed as input classifier predicts scene. predicted then preset user's choice subset. tuned selection. forest prediction 100%. This approach has great potential tedious parameters by allowing individuals participate life adjusting