Forum

Kinect multitouch

Hello everyone :)
I was fiddling around on my computer and found something i got a while ago :
https://youtu.be/MdiCVWC-LeY

The first thing was I installed OpenNI drivers and Yes it works … well enough to toy around with it. Problem came when i mixed up OpenNI and MS-SDK via the bridge, it works but way worse.

Next I dug up in the Kinect Toolkit, i believe the answer is there (somewhere ) but i seem to be too clueless as to how to achieve it.

I was wondering if this can be done in vvvv ?

If anyone has some ideas to point me or the knowledge to do it … Might also be a good addition the the Kinect Toolkit for everyone to use.

Don’t get me wrong i want to learn how to do it but i am too clueless.

So if anyone has some input i will be very happy to receive it :)

Greetings :)

BTW just so it is clear. I am unemployed, sick at home (for months now) and bored out of my scull. All i got is Arduino Mega, Kinect and PS3 eye camera. So i am looking for something to toy with :)

Hi Synth,
as you said, the solution might be somewhere in the Kinect Toolkit. If you’re machine is not able to run DX11.1, try old release of Kinect Hitboxes

[sites/all/modules/general/pubdlcnt/pubdlcnt.php?file=[/sites/default/files/uploads/KinectVirtualHitBoxes_v2.42b_public.zip&nid=94679](https://vvvv.org/sites/all/modules/general/pubdlcnt/pubdlcnt.php?file=((/sites/default/files/uploads/KinectVirtualHitBoxes_v2.42b_public.zip&nid=94679)](https://vvvv.org/documentation//sites/default/files/uploads/KinectVirtualHitBoxes_v2.42b_public.zip&nid=94679](https://vvvv.org/sites/all/modules/general/pubdlcnt/pubdlcnt.php?file=((/sites/default/files/uploads/KinectVirtualHitBoxes_v2.42b_public.zip&nid=94679))
( Contribution is here: kinect-hitboxes-dx11 )

Align Kinect and coordinate system with your “touch” area, set box on top of it,
node CollisionDetection does have output pin “OutputPoint” which gives you a spread of XYZ vectors of the points within defined area. Make a new render target, let the point render as sprites from a top down orthographic view, turn it into a video with AsVideo and track it with Contour node. Old, ugly and CPU heavy compared to DX11 approaches, but it will do what you need :). Pro tip: You can map the distance/Y coordinate of the point to size and or brigthness so you can respond to the distance from the touch area as well.

Get well soon!

A.

Hello and thanks for the input.

So … i need to move to Windows 8.1 for DX11.1 … damn …
I have an nVidia GTX760 card and it supports DX11 for sure but it seems Windows 7 does not support DX11.1

Strangely enough when starting anything from the Kinect Toolbox i get no errors or red nodes so it looks like it is supported.

Windows 7 should support DX11.1: you need SP1 and something you’ll find here:
http://www.microsoft.com/en-us/download/details.aspx?id=36805
See this:
https://support.microsoft.com/en-us/kb/179113

Ok , got it. All is good with my hardware and software. I got misled by this: https://msdn.microsoft.com/en-us/library/jj863687.aspx

Playing with the Kinect Toolkit for now as i am long way from understanding how anything works… Seems this is a bit too ambitions for my level of understanding in VVVV programming :(

i will try to use NecTouch for now as this is the only way i can play around with MT stuff.

Thanks for the input all :)

Hi synth@
If you don’t want to resolve everything in VVVV, you can try LKB, which needs OpenNI. Looks pro, and seems to be quite simple
https://code.google.com/p/lkb-kinect-bundle/downloads/list
Have fun!

Hi Rogalag,

i tested that one too but i think NecTouch works better. I just wanted to try and do it in vvvv and also dont want to destroy my computer by installing both OpenNI and MS-SDK. Soemhow nothing works properly after i install them both.

Thanks for the ling tho :)