Facia has launched DeepLiveness, a product that combines 3D liveness detection with deepfake detection in a single verification session. The solution targets the convergence of AI-generated synthetic faces and injection attacks that have rendered standalone liveness checks insufficient for high-assurance identity verification.
DeepLiveness layers deepfake detection into the liveness check without requiring separate calls or additional steps from the end user. The system operates across two capture modes: live camera sessions for real-time video capture, and static image uploads, which address synthetic AI-generated selfie images submitted to asynchronous verification pipelines, a fraud vector that standard liveness checks cannot address.
The deepfake detection layer runs more than 10 detection models in parallel alongside the 3D liveness check. Facia reports the system detects more than 53 spoof attack types, including deepfakes, 3D masks, 3D-printed models, and digital injection of pre-generated media. The company’s deepfake detection engine achieved 100 percent accuracy on Meta’s deepfake challenge dataset, which covers more than 2,100 videos. Real-world deepfake detection accuracy is reported at 89.91 percent.
Facia’s latest algorithm update reduced False Acceptance Rate to 0.06 percent and False Rejection Rate to 0.03 percent. The platform maintains ISO/IEC 30107-3 iBeta Level 2 certification for presentation attack detection. DeepLiveness is available via API and SDK with sub-second processing latency, and is positioned for identity verification providers, KYC and AML platforms, and financial services firms facing AI-era fraud threats at onboarding and authentication.
The product addresses a threat documented in real-world intelligence, including sophisticated iOS video injection tools enabling deepfake attacks that replace live camera feeds with synthetic video to defeat standard liveness checks. DeepLiveness enters a competitive field, joining a wave of vendors combining liveness and injection attack defense in a single offering, including Jumio’s premium liveness tier. Facia’s approach responds to the deepfake threat to liveness detection systems, which has driven demand for multi-model architectures checking both presentation and injection attack vectors simultaneously.
Gartner has forecast that by 2026, 30 percent of enterprises will no longer consider standard identity verification solutions reliable in isolation.
Sources: Facia, Marketers Media, Facia Blog
–
By the ID Tech Editorial Team
