Real-time rPPG · Anti-Spoof · Web

Contactless Health Vitals & Liveness Detection
Powered by your webcam.

This demo turns a simple camera feed into a signal-aware engine: health vitals via rPPG and a liveliness gate to catch spoofed photos and replayed videos.

Start Vitals Scan❤️Run Liveliness Test🎭
Demo Ready
Live webcam & spoof stress
Signal-aware
Motion, blink & rPPG quality
Liveliness Engine🛡️
Fake Video & Spoof Detection

Focused workflow for “real vs fake” checks using ONNX-based liveness, motion patterns, and blink behaviour.

Liveness score0.00 – 1.00
Static / DeepfakeREAL LIVE
45s Vitals Window📊
Health Vitals Monitor

Forehead ROI rPPG pipeline estimating heart rate, respiratory rate, SpO₂ and more with signal-quality smoothing.

  • Heart Rate72 BPM
  • SpO₂98 %
  • Signal QualityGood
⚙️Pipeline Overview

Face mesh → ROI extraction → rPPG signal → vitals estimation → liveness fusion (ONNX + motion + blinks).

MediaPipe Face MeshrPPG (CHROM)ONNX LivenessMotion & Blink Fusion

Under the hood: signal-aware liveness & vitals

One capture session feeds two flows: a short liveness gate and a longer vitals window — no extra hardware required.

1
Capture & Face Tracking

Face detection / mesh keeps the subject centered, tracks micro-motion and blinks, and defines a clean ROI for rPPG.

2
Liveliness Gate (Short Window)

First few seconds focus on liveness: ONNX model output is fused with blink frequency and micro-motion to tag the feed as real, static or replayed.

3
Vitals Window & Analytics

Once liveness is green-lit, a longer rPPG window estimates heart rate, respiratory rate, SpO₂ and derived metrics with medical-range smoothing and signal-quality checks.