I was asked to edit a number of videos for the Museum of Science’s AI exhibit to demonstrate how AI is being used in different industries. While most of the videos were fairly straightforward a few required some movie (and coding) magic.
I used AE’s facial motion tracking to illustrate how AI proctors detect cheating with facial recognition.
Trying to use After Effects’ motion tracking to imitate pose estimation was still a painstaking process - so to speed things up I went straight to the source. I found an online class to learn the basics of using pose estimation in python and got to work downloading and setting up Pycharm. I was able to follow along with the guide and create the wireframe skeletons, doing days of work in a matter of seconds.
The code used to create the skeleton.
