This project aims to help people who have become voiceless as a result of laryngeal surgery, for example, to regain their own natural-sounding voice. The non-invasive sensor technology can detect and characterize silent speech movements. The latest deep neural networks convert these signals directly into audible speech. The technology developed by Altavo is “non-invasive”: The sensors are attached to the surface of the skin and do not need to be inserted into the body.
After losing their voice completely – as a result of laryngeal cancer surgery, for example – those affected suffer from massive restrictions in their quality of life, ability to communicate and social participation. In Germany and the USA alone, this affects over 50,000 people. Partial or complete loss of phonation can also occur in other clinical pictures. The loss of spoken communication makes medical care more difficult, leads to social isolation, psychological stress and significantly increases rehabilitation costs. Existing substitute voices have so far offered only an inadequate solution: they usually sound unnatural, are difficult to understand and are associated with considerable social stigma and functional disadvantages.
Given this background, the Dresden-based MedTech start-up Altavo is pursuing the goal of developing a novel, non-invasive system for silent speech recognition. Technologically, the system is based on a combination of highly sensitive radar sensors and artificial intelligence, which can be used to detect and decode silent articulation movements and convert them into synthetic speech.
For patients, medical professionals and the environment of those affected, this offers the prospect of significant progress in voice rehabilitation and participation.
Rudolf von Bünau I Altavo GmbH
- Dirk Plettemeier | TUD Dresden University of Technology
- Dr.-Ing. Peter Birkholz | TUD Dresden University of Technology
- Dr.-Ing. Andreas Nocke | TUD Dresden University of Technology
