mixing extracting dancing and skating humans in video with mask-rcnn with a face detected background
together it might be a bit too much? but I'm getting into the idea of using yourself as input, and I'm starting to put the tools together to do that in a way that makes more sense.
sketch video generated from following commands
then, I hooked up a webcam and ran the dlib model to detect faces. that was described in day 010.
the sketch assumes that those directories exist, and that you've checked out the entire
$ python3 sketch-environment.py circleSkate.py 1 -r 600 $ cd imageseq/ $ ffmpeg -r 30 -f image2 -i %*.jpg day15.mp4