| Autor: |
Vanjari , Hrishikesh B., Kolte , Mahesh T. |
| Jazyk: |
angličtina |
| Rok vydání: |
2023 |
| Předmět: |
|
| Zdroj: |
International Journal of Intelligent Systems and Applications in Engineering; Vol. 11 No. 7s (2023): Advancements in Machine Learning for Computer Science and Decision Support Systems; 138-157 |
| ISSN: |
2147-6799 |
| Popis: |
Hearing aids provide the necessary amplification for successful rehabilitation of hearing-impaired persons. It becomes very challenging for hearing aid devicesto attain close to normal hearing. This research suggests a method for improving communication for hearing-impaired people utilizing a combination of three strategies: noise estimation based integrated based gain function, adapted compressive sensing, and listener preference-based customization gain function. Use of integrated gain function and adapted compressive sensing helps to reduce the noise distortion. Use of the customization gain function allows for enhancing the noise-removed speech to the comfort level of the listener. It is achieved by shaping the frequency and amplitude of signal. The overall objective is to enhance quality (noise suppression) and intelligibility (perception) of speech. Performance of proposed solutionis testedagainst noise at variousSNR. Results are compared with existing works to established speech quality metrics. The proposed solution is able to attain about 40% improvement in noise quality and 70% reduction in processing time compared to existing works. |
| Databáze: |
OpenAIRE |
| Externí odkaz: |
|
