How Blood Flow Exposes Deepfakes
From video capture to liveness verdict — how rPPG turns every camera into a deepfake detector
Simple steps to get started
Add Circadify to your existing identity verification or authentication flow with a lightweight SDK. Integration requires no specialized hardware — the SDK works with any standard RGB camera on mobile devices, web browsers, or kiosk systems your users already interact with.
When a user enters your verification flow, the Circadify SDK captures a brief segment of facial video — typically a few seconds. The user experiences zero additional friction. No blinking challenges, no head movements requested. The capture happens passively during the existing selfie or video step.
On-device algorithms analyze the captured frames for micro-color changes caused by blood pulsing through facial capillaries. Circadify isolates these hemodynamic signals from ambient noise, lighting variations, and compression artifacts — producing a blood volume pulse waveform that only living tissue can generate.
Circadify delivers a definitive liveness result: genuine blood flow detected, or not. Deepfakes, printed photos, screen replays, 3D masks, and virtual camera injections all fail the hemodynamic test. The verdict integrates directly into your decisioning pipeline — approve, escalate, or reject.
See rPPG Liveness Detection in Action
Request an enterprise demo to see how Circadify catches deepfakes that bypass traditional anti-spoofing — using the one biometric signal synthetic media cannot replicate.