Microsoft Research had early implementation for this kind of visual sign language translation with Kinect & ASL recognition in 2013 or so. I expect that with the death of Kinect in the market it stayed lab-bound.