Automating drawings using OpenCV image processing

There countless ways to automate Creo, starting with templates, relations, mapkeys, trail files, the Creo API etc. all well known, used and documented. But there is one concept which I use, but hardly ever read about, and that is to use image processing with Creo. And by image processing I mean, taking a screenshot of your desktop screen, analyse it with OpenCV and/or Pytorch and then click for example on a feature. So the process would be something like this:

image processing > trail file/mapkey > image processing > trail file/mapkey > image processing > trail file/mapkey ....

Admittedly it is a rather silly approach, since this has the worst performance you can imagine. Originally it started as one of my isItPossibleInCreoProjects, but still I use at work regularly. Since I couldn't find much on this topic, I made a short video to demonstrate it.