An improved multi-band spectral subtraction technique designed for cochlear implant users

In this paper, we propose a novel approach to reduce noise in cochlear implants by utilizing an enhanced multi-band spectral subtraction method. The main objective of this work is to optimize the subtraction factors used in the spectral subtraction process to improve the overall performance of the stimulation strategy in cochlear implants. These implants are designed to decompose incoming audio signals into multiple sub-bands for further processing. However, one of the persistent challenges in cochlear implant technology is maintaining speech intelligibility in noisy environments. To address this issue, our approach involves adapting the subtraction factors dynamically based on common noise patterns observed in real-world scenarios. To assess the effectiveness of this method, we conducted extensive simulations using noisy sentences corrupted by eight distinct types of environmental noise, including street noise, machinery, and crowd chatter. Objective evaluations were performed using established metrics for speech intelligibility and audio quality. The results demonstrate that our proposed method significantly enhances the signal-to-noise ratio, ensuring improved clarity of speech for users of cochlear implants. This approach is highly promising, as evidenced by the encouraging results obtained in our simulations. By enabling more precise signal processing, it contributes to better speech intelligibility and overall user satisfaction.