Cornell Researchers Develop AI Ring for Real-Time ASL Fingerspelling Translation
The SpellRing prototype uses micro-sonar technology and deep learning to convert ASL fingerspelling into text with high accuracy, aiming to enhance accessibility for the deaf and hard-of-hearing community.
- SpellRing translates American Sign Language (ASL) fingerspelling into text in real time using micro-sonar technology, a gyroscope, and a deep-learning algorithm.
- The device, currently in its prototype phase, achieved an accuracy rate of 82-92% during tests with 20 participants who fingerspelled over 20,000 words.
- Designed to be compact and affordable, the 3D-printed prototype costs approximately $30, with potential for further reductions in mass production.
- Future development plans include integrating the technology into eyeglasses to enable full ASL translation, capturing upper body movements and facial expressions.
- The project, funded by the National Science Foundation, will be presented at an upcoming computing conference in April 2025.
ScienceDaily 
Digital Trends 
Popular Science 
