syntaxaeror/ipd ? reverse-engineered prompt

Build me a Python app that uses a webcam to recognize people and classify their behavior or actions.

I want a simple setup where I can register faces, train the classifiers on my data, then run a main app that watches the camera and shows live results on screen. It should save any needed data and models in the existing data and models folders, and the code should be split cleanly so the camera, config, face registration, training, and main app are easy to understand.

Make it usable from the command line, with a clear way to register a new face and a separate way to train the behavior classifier, then launch the live demo. If you need to look up current docs or best practices for any libraries, go ahead and do that.