When people started tinkering with the Kinect, one of the possibilities that it opened up were new ways to interact with your computers. Virtual keyboards, 3D projections, gesture-based inputs… You name it, somebody was probably working on it. And we’ve actually featured quite a few of them here on the site, and most have gotten quite an incredible reaction from the Kinect Hacks community.

Today, our featured hack is another unique spin on relaying information to computers via gestures. The interesting thing about this hact is it converts those gestures into voice outputs. Here’s a description of the system from its creator,¬†Harishankar Narayanan:

“I was working on some complex gesture recognition system that can recognize numbers. The whole system was built with Embedded systems in mind, so its just c++ and no¬†dependencies. The system can recognize any pre-trained gestures with very good accuracy.”

It’s a simple idea but one that has some potentially great applications especially if combined with other Kinect-based data input systems. It’s really exciting that we’re inching closer and closer to the kind of computers we saw in movies like Minority Report and Iron Man.


Visit Project Website


  1. I just want to know whether it is possible to identify a facial gesture (for an example, a nod gesture) from the user and perform a task accordingly.
    Thank you.