Photoplethysmography

Photoplethysmography studies how light-based pulse signals can reveal heart rate, breathing, blood pressure, oxygen saturation, and cardiac conditions from wearable sensors or facial video. Methods range from filtering and physiological feature extraction to CNNs, Transformers, state-space models, and self-supervised or multimodal learning with ECG and clinical records. Research evaluates these approaches on tasks such as heart-rate tracking, cuffless blood-pressure estimation, atrial-fibrillation detection, and remote PPG under motion, lighting, device, and population changes. A central challenge is whether strong benchmark results generalize to individual people and real-world conditions, making subject-independent testing, uncertainty calibration, and careful checks for data leakage essential.