Anti-Spoofing Mobile Replay Dataset - 3,000+

The Anti-Spoofing Mobile Replay Dataset - 3,000+ is a comprehensive collection designed to enhance biometric authentication systems by addressing replay attacks. This dataset comprises over 3,000 replay attack videos sourced from 1,500 unique individuals, each contributing a selfie followed by a replay video displayed on various smartphones. The replay videos, ranging from 10 to 40 seconds in duration, were captured under diverse lighting conditions and zoom effects to simulate real-world scenarios.
The dataset adheres to ISO 30107-3 standards for Presentation Attack Detection and aligns with FIDO Level B specifications, ensuring its applicability in developing and testing anti-spoofing algorithms. It serves as a valuable resource for training AI models to distinguish between genuine and spoofed biometric data, thereby improving the security and reliability of facial recognition systems. Researchers and developers can utilize this dataset to advance the field of biometric security and contribute to the development of more robust authentication technologies.

反欺骗移动重放数据集-3000+是一个全面的数据集，旨在通过应对重放攻击来增强生物特征认证系统。该数据集包含超过3000个重放攻击视频，来源于1500名独特个体，每人提供一张自拍照，随后是在不同智能手机上显示的重放视频。重放视频时长从10到40秒不等，在多样的光照条件和变焦效果下捕获，以模拟真实场景。
该数据集遵循呈现攻击检测（Presentation Attack Detection）的ISO 30107-3标准，并符合FIDO B级规范，确保其在开发和测试反欺骗算法中的适用性。它是训练AI模型区分真实与伪造生物特征数据的宝贵资源，从而提升人脸识别系统的安全性和可靠性。研究人员和开发人员可利用该数据集推动生物特征安全领域的发展，并助力更稳健的认证技术的研发。