Vital Signs Monitoring (VSM)
Cardiac activity is measured by medical professionals to monitor patients’ health and assist clinical diagnosis. The conventional contact-based monitoring methods, i.e., electrocardiogram (ECG) and photoplethysmography (PPG), are somewhat obtrusive and may cause skin-irritation in sensitive subjects (e.g., skin-damaged patients, neonates). In contrast, camera-based vital signs monitoring triggers a growing interest for non-invasive and non-obtrusive healthcare monitoring.
This project aims to measure various physiological signals from the human face and body using cameras. Supervised and unsupervised machine learning methods are used to extract human vital signs (e.g., heart rate, breathing rate, blood oxygenation, pulse transit time, heart rate variability) from everyday images and videos. These techniques can be used in various camera-based applications that can directly improve people’s quality of life (e.g., tele-health monitoring, neonatal care and fitness tracking).
No project website available.