I have already talked to a few people about that, because I have gotten the Kinect part going. The depth image and getting smooth coordinates from decent gesture recognition is not the problem. A bit more tricky is the inter-process-communication. I did it via a throttled http stream that gets sent from one application (Kinect reading) to the second one (display). A crude version (sans gesture recognition but with process communication) is available on my dropbox.
http://dl.dropbox.com/u/30662912/interprocess_comm.m4v You see the hand position (x,y) and z depth (color) transferred into a client window. So the mouse interrupt handler or whatever would need to read those from a stream. It seems to me useful not for bona fide building for public exhibit displays of molecules, and that is what I intend to use the Kinect for. Resolution is not an issue and smoothing and gesture recognition are also done. The most difficult part was to get the openNI drivers to work with the X-box Kinect, with the usual 32/64 bit hassles etc. ( I am not using the windows Kinect which came out later and not the MS Kinect SDK). I use Processing for the Kinect control. So if somebody is seriously interested in making the basic display program part (coot? Pymol? Other?) work, we can try to crowdsource it or ask the IUCr for some money for IYCr2014 purposes. BR From: Mailing list for users of COOT Crystallographic Software [mailto:[email protected]] On Behalf Of Sebastiano Pasqualato Sent: Monday, May 21, 2012 5:01 AM To: [email protected] Subject: [ot]: will we move structures with our hands? Hi guys, wouldn't it be nice to see this working with COOT? http://www.theverge.com/2012/5/21/3033634/leap-3d-motion-control-system-vide o ciao, s -- Sebastiano Pasqualato, PhD Crystallography Unit Department of Experimental Oncology European Institute of Oncology IFOM-IEO Campus via Adamello, 16 20139 - Milano Italy tel +39 02 9437 5167 fax +39 02 9437 5990
