This program helps experiment with Posy's motion extraction method. Posy proposed a simple algorithm to extract motion from a video. He shows off some pretty trippy results. If you would like to try this for yourself and don't have all the video processing programs that Posy uses, here is a version that you can run in Python.
The code runs on Python 3, using Numy and OpenCV as dependencies. You will need a video that you want to experiment with.
$ pip3 install python-opencv numpy
$ python3 posy-motion-extraction.py some-video.mp4Substitute some-video.mp4 with the file name of the video you would like to
work on.
The slider along the top of the main window allows you to control the time shift (in frames) of the motion extraction.
Press q to exit the program.
This is a quick implementation of video motion extraction, using the method proposed by Posy in his video on Motion Extraction.
It was actually Steve Mould who got me interested in video motion amplification, but I never got round to anything like experimental implementation. The math behind Eulerian Video Magnification was always a bit daunting. When I ran into Posy's video, making this proof of concept was quick and easy.
Just to be clear: the code in this repository does not implement Eulerian, only the simplified version that Posy proposes.
Hope this helps.