Printed 2D Mask Attacks Dataset

The Printed 2D Mask Attacks Dataset provides over 4,800 high-quality videos specifically designed to help researchers, AI engineers, and biometric security providers develop advanced liveness detection and anti-spoofing solutions. This dataset enables machine learning models to learn the subtle differences between genuine human interactions and fraudulent presentation attacks involving printed facial masks.

Dataset Overview

The Printed 2D Mask Attacks Dataset contains more than 4,800 videos featuring individuals wearing or holding printed 2D face masks in front of cameras to simulate presentation attacks against facial recognition systems.

The videos were captured using five different recording devices across diverse environments, helping AI models learn how spoofing attempts appear under varying real-world conditions.

Each sample has been carefully designed to replicate common biometric fraud scenarios encountered by modern facial recognition platforms.

Why Face Anti-Spoofing Matters

Facial recognition systems are now widely used for:

• Digital Identity Verification
• Banking Authentication
• Mobile Device Security
• Access Control Systems
• Border Security
• Employee Verification
• Remote Customer Onboarding

However, these systems remain vulnerable to presentation attacks if they cannot determine whether a real person is physically present.

Dataset Features

Large-Scale Video Collection

The dataset contains more than 4,800 videos, providing sufficient training samples for robust machine learning and deep learning applications.

Realistic Spoofing Scenarios

Videos simulate actual attack methods commonly used to bypass facial recognition systems.

Diverse Recording Conditions

Data includes variations in:

• Backgrounds
• Lighting Conditions
• Camera Angles
• Device Types
• Human Subjects
• Facial Characteristics

Applications Across Industries

Biometric Authentication Systems

Develop stronger facial verification platforms resistant to spoofing attacks.

Banking & FinTech

Protect customer accounts from identity fraud and unauthorized access.

Digital Identity Verification

Improve remote onboarding and Know Your Customer (KYC) processes.

Government & Border Security

Support secure identity validation and access control programs.

Benefits for AI & Security Teams

Organizations can use this dataset to:

 Train advanced anti-spoofing algorithms

 Improve liveness detection accuracy

 Reduce biometric fraud risks

 Enhance facial recognition security

Supporting iBeta Certification Readiness

As biometric authentication adoption grows, certification standards have become increasingly important. Organizations seeking iBeta Level 1 and Level 2 compliance require robust anti-spoofing capabilities capable of resisting presentation attacks.

This dataset provides valuable training data for developing systems that can meet strict security performance requirements and improve resistance against real-world attack attempts.

Conclusion

The Printed 2D Mask Attacks Dataset is a powerful resource for organizations developing facial recognition, liveness detection, and biometric security solutions. With more than 4,800 videos featuring realistic spoofing scenarios captured across diverse environments, the dataset provides the data required to build robust anti-spoofing systems capable of defending against presentation attacks.

Whether you’re improving facial authentication accuracy, preparing for iBeta certification, or developing next-generation biometric security solutions, this dataset delivers the real-world attack data needed to strengthen AI-powered identity verification systems