I see projects like this pop up from time to time. This might be the first one I've seen that is hardware. Usually they are tracking and/or pattern recognition.<p>In any case, while some rudimentary communication might be possible, sign languages tend to have a large improv component, plus facial expressions and body posture all contribute to meaning.<p><i></i>Frighteningly ambitious opportunity<i></i>: Capture <i>all</i> of the signer's output (hands, face, body) and correctly translate <i>that</i> into English.