Real-time implementation of cochlear implant speech processing pipeline on smartphones

Area of Science:

• Biomedical Engineering
• Signal Processing
• Human-Computer Interaction

Background:

• Cochlear implants require advanced signal processing to improve speech understanding in noisy environments.
• Previous implementations of adaptive speech processing pipelines were limited to specialized hardware like PDAs.
• Smartphone ubiquity presents an opportunity for more accessible and adaptable assistive listening technologies.

Purpose of the Study:

• To implement and evaluate an adaptive speech processing pipeline for cochlear implants on a smartphone platform.
• To enable real-time background noise classification and automated noise suppression tuning.
• To assess the feasibility and performance of smartphone-based adaptive audio processing for hearing assistance.

Main Methods:

• Developed an adaptive speech processing pipeline integrating noise classification and noise suppression algorithms.
• Optimized the pipeline for real-time execution on a smartphone, adapting from a previous PDA implementation.
• Conducted performance evaluations measuring real-time processing capabilities and noise suppression effectiveness.

Main Results:

• Successfully implemented a real-time adaptive speech processing pipeline on a smartphone.
• Demonstrated the pipeline's ability to classify background noise environments in real-time.
• Achieved automated, environment-specific tuning of the noise suppression component, showing effective noise reduction.

Conclusions:

• Smartphone platforms are viable for real-time adaptive speech processing in cochlear implant applications.
• The developed pipeline offers a pathway to more accessible and personalized hearing assistance technology.
• Real-time noise classification and suppression on smartphones can significantly enhance cochlear implant user experience.