Optimizing Deep Neural Networks for EEG-Based Speech Recognition: A Multimodal Approach to Assistive Communication
A deep learning-based multimodal framework that combines EEG and audio signals to decode spoken phrases. By integrating neural activity with speech audio, the system improves recognition for individuals with neurological impairments, such as Parkinson's disease, ALS, or cerebral palsy, and enhances robustness in noisy environments. The framework...
Published: 3/10/2026
|
Inventor(s): Anarghya Das, Puru Soni, Wenyao Xu
Keywords(s): Assistive Technology, Technologies
Category(s): Campus > University at Buffalo, Technology Classifications > Artificial Intelligence, Technology Classifications > Medical Devices
|