Why the study?
Conventional remote photoplethysmography methods for noncontact heart rate measurement easily degenerate due to noise interference.
Population
Facial videos from public databases including MAHNOB-HCI
Comparison
Novel CNN-based rPPG method vs other typical rPPG methods
Design
Algorithm development and validation study
Key result
A novel CNN-based remote photoplethysmography method for facial videos achieved a mean absolute error of 5.98 beats per minute and a mean error rate of 7.97% in cross-database testing.
Authors
Loading...
May support contactless HR monitoring in select settings; leaves open prospective clinical validation before adoption.
A novel CNN-based remote photoplethysmography method demonstrates improved accuracy for noncontact heart rate estimation from facial videos compared to conventional methods.
Song et al. (2020) studied Heart rate estimation. CNN-based remote photoplethysmography (rPPG) vs. Conventional and other typical rPPG methods was evaluated on Mean absolute error (beats per minute) and mean error rate percentage. A novel CNN-based remote photoplethysmography method for facial videos achieved a mean absolute error of 5.98 beats per minute and a mean error rate of 7.97% in cross-database testing.