Video Injection Detection
Controls that detect injected, replayed, or pre-recorded video streams used to bypass selfie and liveness checks during remote identity verification.
Overview
Video injection detection focuses on identifying attempts to feed a verification system a manipulated or pre-recorded stream (for example, via virtual cameras, screen replays, or intermediaries) instead of a live capture from the user’s device camera.
Typical signals and controls
- Capture pipeline integrity checks (e.g., detecting virtual camera sources).
- Consistency checks across frames and metadata (timing, encoding artifacts).
- Challenge-response or motion prompts combined with PAD/liveness.
Where it shows up
- Remote onboarding / selfie identity verification
- Account recovery re-verification
- High-risk transaction step-up checks