Revolutionizing Healthcare with NAM Signal and EEG Signal Based Silent Speech Recognition Models

This paper introduces a sophisticated system that translates non-audible murmur (NAM) and electroencephalogram (EEG) signals into text. It is intended for patients who cannot speak out loud, e.g., ICU patients. NAM records minute vocal cord movements, and EEG records brain activity associated with speech. By combining these signals, the system improves silent speech interpretation using deep learning models such as CNNs and RNNs. Another optimization method, SSO, also adjusts parameters to enhance accuracy and efficiency. With a 96% accuracy rate, this system outperforms conventional speech recognition techniques. It provides an essential communication tool for people with severe speech disabilities, especially in healthcare, enhancing their communication capacity.