Facia, a leader in biometric identity verification, today announces DeepLiveness, an upgrade to conventional liveness detection built for the generative AI era.
Traditional liveness detection confirms proof of presence through a live camera feed. DeepLiveness goes further by verifying that the face is not AI-generated and by performing a full authenticity check on static image uploads submitted to the system. Both checks run together in under 1 second, elevating accuracy while unlocking new use cases where an image upload, rather than a live session, serves as the entry point.
Standard liveness was built to defeat printed photos and video replays, not AI-generated faces that pass motion checks without a real human, or digital injection attacks where a synthetic feed is inserted before the check runs. Attackers now deploy both at scale across KYC onboarding and remote authentication. Confirming presence without confirming face authenticity is no longer sufficient.
DeepLiveness combines Facia’s 3D liveness verification with its deepfake detection engine into a single unified check with no added latency. Core technical capabilities include:
- RGB Texture Analysis at 95% signal coverage, identifying synthetic rendering artifacts invisible to standard liveness.
- Motion Analysis operating at 91%, catching face-swap and injection attacks through micro-movement anomaly detection.
- Depth Mapping: at 92% signal strength, verifying 3D facial geometry to block injection and flat-image attacks.
- Neural Anomaly Detection: scoring 96% signal strength, identifying generative model signatures and AI-synthesized identity patterns in real time. Together, all four signals deliver 100% detection coverage, presence confirmed, and face verified as real, with synthetic identities flagged before they reach any downstream system.
- iBeta Level 2 Compliance maintained across the combined verification layer, with 0% Attack Presentation Classification Error Rate (APCER), zero successful presentation attacks on Android and iOS, confirmed by iBeta Quality Assurance, a NIST-accredited testing laboratory. DeepLiveness records a False Acceptance Rate (FAR) of 0.06% and a False Rejection Rate (FRR) of 0.03%, reducing both fraud risk and legitimate user friction across KYC onboarding and remote authentication workflows.
AI-driven identity fraud is accelerating across fintech, banking, and iGaming. Regulators across the EU, UK, and APAC are tightening assurance requirements in response. DeepLiveness closes that gap at the verification layer.
“Standard liveness was built for printed photos. DeepLiveness is built for a world where attackers deploy AI-generated faces and virtual camera injections across four signal dimensions simultaneously, all in under one second. That is the new baseline.” – Daniyal Chugtai, CTO, Facia.
DeepLiveness is available now through Facia’s existing SDK and REST API. No new integration required.
About Facia: Facia is committed to setting a new standard in identity verification by advancing liveness detection and deepfake detection solutions that keep pace with the rapidly evolving threat of AI-generated fraud.