Keywords: Computer Vision, Eulerian Video Magnification (EVM), Unobtrusive Human Sensing
This project was done individually, during my first semester at Carnegie Mellon, as a part of the “Computer Vision” course.
Past research shows that blood flow in the facial region increases when a person experiences pain. The goal of this project was to estimate pain by measuring blood flow. To this end, I implemented Eulerian Video Magnification in Matlab to magnify color changes in videos of people with shoulder injuries experiencing pain. Preliminary analysis of blood flow measured using this method and annotated pain levels shows a relation between the two.
This project highlights the potential of using EVM to unobtrusively detect physiological phenomena like pain, and magnify color and motion, such as exaggerating facial expressions by magnifying the movement of facial landmarks.