This is me calibrating the Kinect with the prototype. In the video, I'm testing whether the Kinect can detect the pushed points of the screen.
The code used was based on D. Shiffman's tutorials on interaction and thresholds with Kinect and Processing.
Figured out the proper dimensions and distance needed for the final layout
Might have figured out the detection distance thresholds needed for the final set-up
The ball shape in the code is created by averaging the locations of each point within the threshold range. It's not precise enough because there are tiny points being detected by the Kinect.
Based on the issue above, I need to make the Kinect's input more precise. I may need to just remove any stray objects within the area or somehow "crop" any thing outside of the frame.
Also, the Kinect I'm using is limited to the canvas size you see in the video. I may need to position the Kinect 90º to detect the full height of the frame. I need to figure out how to deal with that.
Finally, I need to compress the dimensions of the set-up and make it more compact.