Search papers, labs, and topics across Lattice.
The authors investigate whether deep remote photoplethysmography (rPPG) models can be trained on pseudo-labels extracted from classical unsupervised signal-processing methods rather than hardware-synchronized contact sensor data. They find that pseudo-label supervision actually outperforms ground-truth sensor training on datasets with imperfect temporal alignment and matches supervised performance on well-synchronized data in within-dataset regimes. While cross-dataset generalization slightly favors supervised models, simple filtering of outlier labels closes the gap, proving that algorithmic labeling can substitute for fragile hardware setups.
Unsupervised signal-processing pseudo-labels can outperform ground-truth contact sensors when training deep physiological vision models on imperfectly synchronized video.
Heart rate is a critical biomarker of health, and remote photoplethysmography (rPPG) enables its contactless estimation from video data for telemedicine applications. Recent advancements in deep learning based rPPG methods achieve state-of-the-art results, outperforming classical signal-processing methods in complex scenarios. However, deep learning methods depend on datasets with precise synchronization between videos and ground truth signals collected via contact sensors, whereas signal-processing-based methods do not. To address this dependence on labeled datasets, which are labor-intensive to collect, we investigate under which circumstances pseudo-labels extracted using unsupervised signal-processing methods can replace contact sensors labels for training deep learning methods. Our systematic evaluations found that for datasets with imperfect synchronization, the pseudo-label approach outperforms supervised training on contact sensors. For datasets with good synchronization, results are mixed: within-dataset evaluation shows no significant difference between training methods, while cross-dataset evaluation favors supervised training. However, removing a single outlier participant significantly improves the pseudo-label approach's cross-dataset performance, highlighting the importance of label quality. These results demonstrate that signal-processing methods can generate valid training signals for deep learning models, reducing dependency on labor-intensive dataset collection while maintaining competitive performance.
