For Project 1 I put more effort into creating an interesting system and less on its aesthetic output. That said, the code below (while embarrassingly messy at the moment) uses a few different computer vision techniques to track and record people’s position in front of the camera and in turn display who in the past has occupied the same general area as the current viewer. The code also removes the background from the footage being captured so when it is played back to future viewers, only a past figure appears and not the background (if the bg remove is clean). Finally, as a current viewer is looking at the piece, they are also being recorded and their positions saved into the system and thus in advertently contribute to a future viewer’s experience of the piece.
Once again, the code is not perfect and I am still having a hard time accurately recording the face positions, but I think it is a good start. There is also probably a much more efficient way to get what I have now, but I need to learn a little more before I get to that point. Finally, asthetically it could be better, but I decided to get a working system first.
If you would like to take a look at the code, it can be downloaded from this link: