South Korean researchers have built a sign language translator from seven wireless rings, each containing a three-axis accelerometer, the same tiny motion sensor that rotates your phone screen when you tilt it. Worn on selected fingers, the rings track finger orientation and hand movement, send the data to a phone via Bluetooth, and AI converts the gestures into text in real time.
The system, called WRSLT (wirelessly connected, ring-type sign language translator), recognizes 100 American Sign Language words and 100 International Sign Language words with 88% accuracy. That accuracy was measured on unseen users whose signing data wasn't used to train the AI, which is the test that matters for a real assistive device, not just a controlled lab demo.
Every existing sign language translation tool uses gloves with embedded sensors or wired sensor arrays. They're bulky, highly visible, and don't fit different hand shapes, which limits how long and where anyone can actually wear them. Consumer gloves like BrightSign, which ships a Version 2 this summer, have been working on this problem for years. Seven independent rings sidestep it entirely. Each one sits on a finger without connecting to the others, so hand size and shape don't matter.
The researchers also note the ring architecture could extend beyond sign language into gesture control for augmented reality, touchless device interfaces, and rehabilitation monitoring. Consumer rings from Oura, Samsung, and RingConncurrently track health metrics passively, your heart rate, your sleep, your temperature. RingConn's Gen 3 just shipped haptic feedback that can tap your finger for health alerts, which is the first step toward rings that do more than observe. WRSLT points toward rings that read what your hands are doing on purpose, not just what your body is doing automatically.
The system isn't ready for consumers. The team needs larger datasets, more vocabulary, longer wear tests, and involvement from signing communities before it gets anywhere near a product. 88% accuracy across 100 words is a research result. A useful translator needs thousands of words and accuracy closer to 95% before it holds up in a real conversation. Samsung's Brain Health dementia detection is further along. Clinical trials are already running, but it still hasn't shipped. Both features show where the ring category is headed. Neither is on your finger yet.
Seven rings, no wires, no gloves, and language translation isn't just for spoken or written words anymore.