Silent speech interface

Freitas, J., Teixeira, A., Dias, M. S., & Silva, S. (2017). An Introduction to Silent Speech Interfaces. Springer International Publishing

# AlterEgo

Surface electromyography-based [[zFNrLkCqVyTK22wRfDKDsqy:::Silent speech interface]] project by [[zDaBAuDF7bsCxaRcrAMnRN6:::Arnav Kapur]], in [[zFGxUhgWwhyY43iWCakDHKk:::Pattie Maes]]’s group at [[z46zxKqhtFPVVVPXKFo9vHn:::MIT Media Lab]]. First publication in 2018. Its key innovation (AFAICT) is to bring contemporary deep learning practices to the EMG signal processing.

*   A. Kapur, S. Kapur, and P. Maes, “AlterEgo: A Personalized Wearable Silent Speech Interface.” 23rd International Conference on Intelligent User Interfaces (IUI 2018), pp 43-53, March 5, 2018.
    *   “On a standard digit recognition test, our system achieves a median accuracy of 92%” Very limited vocabularies (20 words).
*   Kapur, A., Sarawgi, U., Wadkins, E., Wu, M., Hollenstein, N., & Maes, P. (n.d.). Non-Invasive Silent Speech Recognition in Multiple Sclerosis with Dysphonia.
    *   Expands to recognizing sentences from a limited set (15 phrases); achieves ~200 bits per minute at a mean accuracy of ~0.8






