Kinect Depth Culling

I bought myself a Kinect sensor this week; I haven’t had much chance to really experiment thus far in 2012, and thought this might be a good place to start. That, and the Kinect is one of the less invasive interactive tracking instruments I can think of. I am likely to be creating interactions for children with autism, so for many reasons, the more I can make the technology invisible, the better.

I toyed with the idea of background removal last year, so revisiting it seemed logical. It didn’t take long for me to find someone who is doing the same thing much, much better…

%CODE1%

If you’re following along with all of that at home, the creator of the above video has kindly provided his patch online. I tried to pull apart the above patch, but realised pretty quickly that it was beyond me. It was also a lot more complicated than what I had in mind, so I tried a relatively simple depth culling version, stealing some of Seldess’ good work…

It’s certainly not yet as robust as Seldess’, but I thought it may be useful for anyone new to this (like me). It might be a helpful start if you’re trying to build your own patch with a Kinect. Here’s the code.