I haven't done a lot of pd systems involving visuals, so I'm unsure of how evolved and documented the video-oriented side of PD is.
I'm in the very early stages of experimenting with what will eventually be an interactive system controlling physical parameters.
Right now, the idea in its core is this: A pd patch reads/scans a pre-recorded or live-input video source and based on what values can be extracted in realtime (or offline), these values will control various stages of the system.
At this stage I am not interested in synthesising visuals with GEM, but to "read" the information of a real-time video source. Something along the lines of movement, colour, contrasts, the ratio of light versus dark in the video, maybe scans of grain, framerate and so forth.
Are there any places I can find documentation on similar systems? I imagine I would have to incorporate a second application of sorts, a program to analyse and output the video data that I could then patch into PD and use as control-values.
Any and all information and suggestions will be appreciated.