Hi,

I hope someone can help.

We have been working on an augmented reality fitting room using Kinect. One of the problems we've found is that whe the person turns around, the Kinect stops tracking the body for a while. I have read some other related topic here and it appears the solution is to use 2 Kinects.

We will build a module that will contain the screen, the kinect and an HD camera, however, if it's two Kinects, how do they need to be arranged? could these two kinects be in front of the person, separated lets say by 3 feet, one besides the other?, or do I need to put the 2nd kinect behind the person?

Thank you very much in advance for your input.

Best,
Carlos Marruffo

Hello

Of course, placing second Kinect behind the person will do a better job for continuous tracking. Because this practically eliminates the cause of tracking errors - the obscurity of body parts while turning.

However, you'll have some problems using 2 Kinects in your app.
If you plan to use our software - be aware that it's not realtime currently, which I suppose is a requirement for you.
If you use Kinect SDK skeletal tracking - though it allows to get raw data from multiple sensors connected to the same PC, the skeletal data is allowed for only one of them. This may change with Kinect 2 SDK, yet to be seen. If that is possible, you will need to merge (or at least choose) skeletal data from both sensors on your own.