Certain Investigation of Various Algorithms to Improvise the Quality of Hearing Aid

Hearing loss, marked by an inability to hear sounds below a 20 dB threshold, often results from dysfunction in the ear’s organs, associated nerves, or the auditory part of the brain. Deaf individuals encounter direct or indirect discrimination in daily life, as highlighted by National Centre for the Deaf data showing a significant employment gap. Only 48% of deaf individuals are employed, compared to 72% of their hearing counterparts. This discrimination leads to missed opportunities and some deaf individuals leaving jobs due to disability-related biases. After careful study of research papers done by scholars and working with low-end and High-end Hearing aid devices, we decide to start with the analysis of audio signals and implementing the different algorithms in computational tools like MATLAB and Python. Subsequently, we found the quality of hearing aid devices depends on three factors SNR, SI-SNR, and PESQ. Working with different algorithms and modifying some, we achieved better quality hearing aids than existing high-end hearing aids using the SepFormer algorithm. The overall systems shows better results on the signal quality improvements to ensure the HA perforamnce. The Metric GAN+based method has the good variations in SI-SNR of 9.42, PESQ of 3.15. Hence, using the proposed model enhance the quality of HA and supports the dumb people. The quality of the HA is analysed using this proposed mehtod with AI enabled technlogy.