Andrew Heinzman
Bioengineers at UCLA are developing a glove that interprets American Sign Language into spoken English through a smartphone app. The system works in real-time at 60 words per minute, with a 98.63% percent accuracy. While far from being a finished product, UCLA’s ASL gloves could become a resource for education or accessibility in the future.
But why a glove? Unlike a camera system, UCLA’s ASL gloves have the potential to be portable or innocuous. They’re lined with thin stretchy sensors and electrically conducting yarn that detects motions and gestures. The gloves connect wirelessly to a phone that, with the right app, can detect and interpret ASL into spoken English on the fly. Also, unlike a camera system, the ASL glove only costs about $50 in parts and could connect to facial sensors to capture expressions that are part of ASL.
The ASL glove is impressive, but it’s still a prototype. It interprets words at 60 words per minute, which is slower than conversation speed for people who are proficient at ASL. Also, it only has a vocabulary of 660 words—about half the vocabulary of Koko, the gorilla.
For all its limitations, the ASL glove has the potential for large-scale production and popularity. UCLA already filed a patent for the glove and hopes to manufacture it for sale once it works faster and has a larger vocabulary.
