Heart Rate Estimation
Overview
Heart rate estimation aims to recover beats per minute from signals such as contact photoplethysmography (PPG), electrocardiograms, camera-based remote PPG, radar, thermal video, and accelerometers. Methods range from filtering and spectral peak detection to deep temporal networks, self-supervised learning, and fusion across sensors; many are tested on public video benchmarks, wearable recordings, and clinical or in-the-wild datasets. Current work targets robustness to motion, lighting, pose, sensor placement, and missing or poorly synchronized signals, with reliability scoring and selective prediction used to trade coverage for accuracy. A central challenge is whether strong benchmark results transfer across people and settings, while remaining fair and interpretable and distinguishing genuine physiological information from artifacts or chance agreement.
Latest papers
31 papers5 papers in the last four weeks, up 67% on the four weeks before. 0.0% of all new papers.