Cornell Researchers Unveil SpellRing for Real-Time ASL Fingerspelling Translation
The prototype wearable device uses AI and micro-sonar to translate ASL fingerspelling into text with up to 92% accuracy, offering a low-cost solution for accessibility.
- SpellRing is a thumb-worn device that translates American Sign Language (ASL) fingerspelling into text in real-time using deep learning, micro-sonar, and gyroscope technologies.
- The prototype achieved an accuracy rate of 82–92% during tests with both novice and experienced ASL users, spelling out approximately 20,000 words.
- Developed by Cornell University researchers, the device costs around $30 to produce, with potential for further cost reductions in mass manufacturing.
- Currently limited to fingerspelling, the team plans to expand SpellRing's capabilities to recognize full ASL gestures and integrate with AR glasses for broader applications.
- SpellRing is being positioned as a fast and accessible text input tool, with potential uses beyond accessibility, such as web searches and smartphone navigation.
ScienceDaily 
