Researchers from POSTECH (Pohang University of Science and Technology, Korea), have developed a wearable AI-powered sensor that can reconstruct natural speech by detecting microscopic movements of the neck muscles and skin — even when no sound is produced by the vocal cords.
How It Works
- The soft, silicone-based “Multiaxial Strain Mapping Sensor” is worn comfortably on the neck.
- It uses a miniature camera and tiny markers to track subtle strain patterns caused by attempted speech (subvocalization).
- An AI algorithm interprets these invisible movement maps to figure out the intended words or sentences.
- The system then synthesizes the speech in the person’s own pre-recorded voice using personalized voice synthesis.
The technology includes smart calibration to correct for slight differences in sensor placement and performs well even in noisy environments.
Who It Helps
This offers new hope for people who have lost their voice due to:
- Vocal cord disorders
- Laryngeal cancer surgery (including full laryngectomy)
- Other speech impairments
It allows them to “speak” silently and have their own voice restored.
Key Advantages
- Non-invasive and hands-free
- Works without any vocal cord vibration
- Suitable for daily use, including noisy workplaces or quiet settings (e.g., libraries or meetings)
Professor Sung-Min Park (POSTECH) stated: “We hope this technology will accelerate the day when patients with speech disorders can reclaim their voices.” The study highlights the sensor’s potential not only for voice restoration but also for silent communication in various professional and everyday scenarios.
