Comparison of Different Facial ROI for Remote Photoplethysmography

The goal of this course project was to compare the performance of select Remote Photoplethysmography algorithms on different facial ROIs. A secondary goal of the project was to analyse the effect of computational imaging on rPPG by varying imaging conditions such as exposure time, aperture and camera ISO.

From various experiments, we determined the lowest RMSE of 8 bpm by predicting the heart rate from the temple at an ISO of 460 using POS (Plane Orthogonal to skin) algorithm