Better face detection: colour images, saves the original frame. Also (with the help of the php script) writes some info to a database: time, location (hardcoded for now) face position and dimensions, folder name, original image name. Everything can be linked back to the original and to everything else.
The next step is to do some after-the-fact analysis of the faces. Some kind of shape recognition so I can start swapping similar faces in the originals.
I would love to make this some kind of collaborative effort, where I try to make a narrative out of the material I’m gathering, while the computer’s trying to do the same thing, using its own criteria. Hopefully we’d work well together some of the time, but also completely disagree at other times. But first I’m going to try and do something similar with sound. That’ll be what I present for class next week.