MyVoice: Myoelectric Vocal Interaction and Communication Engine

• Applicant: Professorin Dr.-Ing. Tanja Schultz
• Subject Area: Human Factors, Ergonomics, Human-Machine Systems ; Image and Language Processing, Computer Graphics and Visualisation, Human Computer Interaction, Ubiquitous and Wearable Computing
• Term: since 2022
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 460884988

Silent Speech processing aims at making speech communication available to persons without a voice, and to people who require covert communication in public places. A teamof two researchers in Switzerland and Germany, with more than 10 years of experience in the field, proposes to substantially advance their state of the art system, making it a truly European technology which in the future can improve the lifes of hundreds of thousands of speech-impaired persons, and offer an efficient means of private and undisturbing communication to the general public. We build on our existing work in Silent Speech processing with state-of-the-art neural network architectures, where bioelectric traces of articulatory movements of a speaker are captured with small, nonintrusive surface electrodes. These movements capture speech even when the acoustic signal is disturbed or not present at all . The project MyVoice we will tackle the key roadblocks which so far have prevented Silent Speech processing from becoming a communication tool and speech prosthesis, namely fast and reliable training and adaptation mechanisms in a human-in-the-loop process which provides instant audible feedback to the user as well as prosody generation and output naturalness to generate prosodic contours that are both semantically plausible and acoustically pleasant in spoken communication. The proposers team is complemented by two non-funded cooperation partners from Spain (Prof. Inmaculada Hernaez, University of the Basque Country and Prof. Jose Gonzalez, University of Granada), who will deploy the MyVoice system in patient studies with laryngectomized patients, as well as counseling on specific technical aspects, as detailed in the research plan.

• DFG Programme: Research Grants
• International Connection: Switzerland
• Cooperation Partner: Privatdozent Dr.-Ing. Michael Wand