Face Anti-Spoofing and Liveness Detection Datasets

A liveness detection dataset is a collection of video or image data used to train AI models that distinguish real human faces from presentation attacks, such as printed photos, display replays, or 3D masks worn by an attacker. Liveness detection datasets are essential for any biometric system that relies on face-based identity verification, particularly in KYC, banking onboarding, and fraud prevention. High-quality datasets cover multiple attack types, demographic diversity, and capture conditions to ensure models generalize beyond lab environments

Axon Labs provides a comprehensive catalog of face liveness detection datasets covering every major attack class: photo print attacks, display replay attacks, silicone masks, latex masks, cloth 3D masks, wrapped 3D masks, high-fidelity rubber masks, and 3D resin masks. Our datasets are organized by iBeta certification level: iBeta Level 1 covers 2D attacks, iBeta Level 2 covers 3D masks, and iBeta Level 3 covers the highest-fidelity mask attacks

The right liveness dataset depends on your target certification level. For iBeta Level 1 - use the iBeta Level 1 Dataset (covers paper print, cutout, and display replay attacks). For iBeta Level 2 - use the iBeta Level 2 Dataset or individual 3D mask datasets (silicone, latex, wrapped 3D, cloth). For iBeta Level 3 - use the iBeta Level 3 Dataset or specialized high-fidelity rubber and 3D resin mask datasets. 21% of companies that passed iBeta certification in 2025 are Axon Labs clients

Selecting a face liveness detection dataset depends on four factors: (1) the attack types you need to defend against: 2D prints, display replays, or 3D masks; (2) the certification level you target: iBeta Level 1, 2, or 3; (3) the scale and demographic diversity your model requires; and (4) whether you need active liveness protocols (head movements, zoom, blinking) or passive detection. Axon Labs datasets cover all four dimensions across the full catalog, with detailed specifications on each individual dataset page

Axon Labs liveness detection datasets cover the full spectrum of presentation attacks tested under ISO/IEC 30107-3: 2D attacks (photo prints, cutouts, paper masks, cylinder-mounted prints, smartphone and PC display replays); 3D attacks (silicone masks, latex masks, cloth textile masks, wrapped 3D paper masks, high-fidelity rubber masks, and custom-manufactured 3D resin masks). This coverage enables training a single model against every attack class encountered in real-world biometric fraud

Yes. All Axon Labs liveness detection datasets are collected with explicit written informed consent from every participant in compliance with GDPR Article 9, which specifically governs the processing of biometric data. All datasets are licensed for commercial use in AI model training, validation, and iBeta certification preparation. Comprehensive compliance documentation, including consent provenance, collection methodology, and legal basis review, is available upon request for regulatory or audit purposes

Yes. Free samples are available for every liveness detection dataset in our catalog, covering photo print attacks, display replays, silicone masks, latex masks, cloth 3D masks, wrapped 3D masks, and high-fidelity rubber and resin masks. Samples allow you to verify attack diversity, video quality, active liveness sequences, and format compatibility with your training pipeline before committing to the full dataset. Submit a sample request, delivery typically takes 1–2 business days