Modding Microsoft’s Kinect is nothing new; people have been designing new and crazy things with it for over a year already. There is even a Kinect for Windows release in February 2012 that is spurring innovation. However, occasionally someone uses it to do some pretty ingenious feats, like George MacKerron, a researcher at University College London.
MAcKerron took the Kinect and made an interactive ‘Depthcam’ that lets the user interact with real-time 3D images on the web. To do this George used OpenKinect Python wrapper along with Autobahn Websockets library to connect the Kinect to a web browser (in this case Chrome). To get around the networks firewall (at UCL), he used a node.js server, which is built upon Chrome’s version of JavaScript’s runtime environment. He then used CoffeScript to connect to the node.js server which translates the data received as a particle system that uses WebGL. As it stands right now, it only works with Google’s Chrome browser and is limited in content when using a mouse to pan and scan the live image, but it’s still pretty impressive all the same. To get a full rundown on George MacKerron’s Depthcam, or get all the code to try it yourself, visit his website.
Cabe