Kinect Patent Says It Can Read Sign Language

Illustration for article titled Kinect Patent Says It Can Read Sign Language

The patent for Microsoft's motion-sensing camera Kinect was released to the public last week, and while we knew most of the stuff contained within, I'd be lying if I said I knew the device could recognise sign language.


But it does. At least, according to the patent it does (things may have changed since, we're checking), as it says that thanks to Kinect's ability to track hand and arm gestures, it can understand somebody signing at it in American Sign Language, and then convert it to either text or spoken word.

A neat feature! It might be totally useless to most of you, but anything that can improve the accessibility of games - and do it as seamlessly as allowing someone to sign at their TV - is a welcome step forward nonetheless.

Kinect patent detailed; American Sign Language supported [SlashGear]


This appears to contradict the DigitalFoundry article I read recently:


"Perhaps the biggest surprise - and disappointment - in viewing the final Kinect spec up against the PrimeSense reference camera is that the resolution of the all-important depth map has been reduced to just 25 per cent of its original size: 320x240 up against 640x480.

This makes the idea of tracking fingers, or any kind of fine detail, in a reliable fashion a complete non-starter, and so what you might call "higher level" interactions with games are going to be challenging to implement: there won't be any virtual keyboards to type on, for example: the precision simply isn't there. "