A research team at Cornell, led by doctoral student Hyunchul Lim, has created SpellRing, an AI-powered ring using micro-sonar to track fingerspelling in American Sign Language (ASL) in real time. Unlike previous bulky devices, SpellRing is compact and designed to cater to the needs of the deaf and hard-of-hearing community. It enables users to input text into computers or smartphones through fingerspelling, which spells out words not represented by specific ASL signs. SpellRing is worn on the thumb and contains a microphone, speaker, and mini gyroscope—all enclosed in a 3D-printed casing the size of a quarter. These components work together to send and receive inaudible sound waves that monitor finger movements.

A proprietary deep-learning algorithm processes the sonar data to recognize fingerspelled letters with an accuracy rate of 82-92% based on testing with over 20, 000 words from both novice and experienced ASL signers. The project, presented at the Association of Computing Machinery’s CHI Conference in Japan, aims to bridge the gap between technology developers and the ASL community. However, researchers noted that recognizing the variations in ASL fingerspelling remains challenging. This project advances from a previous model called Ring-a-Pose and aligns with the SciFi Lab's trajectory of developing sonar-equipped devices for various applications. Future enhancements may include integrating the micro-sonar system into eyeglasses to facilitate more comprehensive ASL translation by capturing upper body movements and facial expressions, which play a significant role in ASL communication. The project has been supported by the National Science Foundation, with contributions from various co-authors in the fields of information science, computer science, and linguistics.