[thelastnode]

This also works with live camera feed because of how OpenCV handles video streams. You can extract frames from the stream and then run the same face detection code on those extracted frames.

This is a great post given that a lot of the OpenCV functions, while powerful, are either missing documentation or have outdated documentation.