What Is [Vision] DepthTracker?
[Vision] DepthTracker appeals to those who cannot afford a full multitouch setup, or who only have one camera and want to test the abilities of relative depth tracking on a minor scale. With features such as threshold, background removal, unique tracking with depth analysis and blob size restrictions, and the obvious ability to experience a natural user interface with a single camera. For even more info, please visit this previous thread.

Why Should I Care?
With the ingenuitive use of simple lighting techniques and a single camera setup, one can easily experience the true potential of a natural user interface without having to buy thousands of dollars worth of special equipment and other devices.

Where can I download it?
See the bottom of this post to download the application via the attachment.

   

Some notes for your consideration...

When the application is first started on windows, it will give a “Memory Access Violation” because windows will ask to allow for the port to be open and firewall permission to be granted to this application. To fix this, simply allow the firewall exception to be made by clicking “Unblock” and run the application again.

Required for application to run:
PS3eye Camera
Light source either behind or in front of the camera (as shown in attached diagram).
A copy of the application via download.

Recommended PC specifications:
1.6 Ghz Dual Core
2 Gb RAM
DirectX 9.0 or above



*** Special thanks to AlexP for his works on getting the PS3eye camera to work under windows! ***

[Vision] DepthTracker by hillbilly is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License.

thank you hillbilly for posting the app. i just tested it out and got really amazing results. is it possible to use two cams like kinect to have more 3D environment ?

Hi again, vahid! I believe the two camera setup you mentioned to be possible, yes, and I think one way to do this is to cross-analyze the depth mesh of one camera with the depth mesh of a second one (nearby the first, of course), and then interpolate the data retrieved to get a better 3D image. I’m thinking of starting some tests now and, hopefully may have something to show in the next post here.

Amazing work, very clever.

This looks really awesome thanks for sharing and looking forward to experimenting with this (esp with some IR)

Thanks everyone! And, Christian, please do experiment, as I am limited on the ability to do such tests, especially with IR. Thanks again!