Researchers at Cornell University have developed SpellRing, an AI-powered ring that translates American Sign Language (ASL) finger spelling into text in real-time.
The device, equipped with micro-sonar technology, tracks the subtle movements of the fingers. The ring, worn on the thumb, emits and receives inaudible sound waves to monitor hand and finger motions, complemented by a built-in mini-gyroscope. A deep learning algorithm processes the micro-sonar data to predict letters described in ASL.
Tested with 20 ASL users, SpellRing achieved an accuracy rate between 82% and 92%, depending on word difficulty. The goal is to provide more diverse and inclusive access to computational resources for the ASL community.