Imagined Speech: Imagined speech refers to the process in which a subject imagines speaking a given word without moving any muscle or sound.
The ability to understand imagined speech will fundamentally change the way we interact with our devices. We’d like to classify the syllables “ba”, “ku” ,“im” and “si” from imagined speech EEG signals. These syllables were selected since they contain no semantic meaning so that classification would be performed on the imagined speech instead of the semantic contribution to imagined speech production.
Classification: Using artificial neural networks, our model is successfully able to classify syllable pairs from the EEG data with over 90 percent accuracy.
Copyright to team members.