Contactless physiological monitoring from a standard webcam.
No wearables. No hardware. No installation. Just a browser.
Live demo → [Deploying Soon]
Scan to try on mobile:
[Coming Soon]
SightBeat uses your webcam to extract real-time health metrics by detecting microscopic color changes in your skin caused by blood flow — a technique called rPPG (Remote Photoplethysmography).
Everything runs locally in your browser. No video is ever transmitted to any server.
| Metric | Method | Scientific Basis |
|---|---|---|
| Heart Rate (BPM) | Green channel rPPG, peak detection | de Haan & Jeanne, 2013 |
| HRV · RMSSD | Inter-peak interval analysis | Task Force ESC/NASPE, 1996 |
| Autonomic Balance | RMSSD sympathetic/parasympathetic index | Thayer et al., 2009 |
| Blink Rate | EAR formula on 6 eye landmarks | Soukupová & Čech, 2016 |
| Fatigue Level | Blink rate + HRV baseline drift composite | Craig et al., 2012 |
| Perfusion Index | R-channel AC/DC ratio | Reisner et al., 2008 |
| Emotional State | HRV + HR trend + activity state fusion | Thayer et al., 2009 |
| CNI Score (0–100) | Weighted composite: HR efficiency + HRV + blink | Shaffer & Ginsberg, 2017 |
| Detailed Inference | Local deterministic multi-metric synthesis | — |
- Node.js 18 or higher
- A webcam (built-in or external)
- Good lighting (natural or warm white light preferred)
- Modern browser: Chrome, Edge, or Firefox
Webcam frame
↓
MediaPipe FaceMesh (478 landmarks)
↓
Forehead ROI extraction → mean R, G, B per frame
↓
CHROM algorithm (chrominance signal separation)
↓
Butterworth bandpass filter (0.75–2.5 Hz)
↓
Kalman filter (motion artifact rejection)
↓
Peak detection → Heart Rate + HRV
↓
Derived metrics: Fatigue, Emotional State, CNI Score