rPPG-Toolbox

Why we included this project

Getting heart rate from an ordinary webcam is harder than it looks. The signal hides in subtle skin-color changes, and reproducing published rPPG results has historically meant reimplementing each method's preprocessing and evaluation by hand. This toolbox bundles the classic unsupervised baselines (CHROM, POS, ICA, GREEN) together with modern supervised neural models in one Python framework, so you can run comparisons without stitching together separate codebases. It also includes public benchmark datasets, augmentation, and standardized metrics, which takes most of the guesswork out of reproducing results. That makes it a practical reference for teams building telehealth or wellness prototypes, and a fair benchmarking foundation for researchers testing new ideas.