Anti-Spoofing Photo Print Attack Dataset

The Anti-Spoofing Photo Print Attack Dataset is a biometric dataset designed for advancing facial liveness detection, particularly to guard against photo print spoof attacks. It comprises over 3,000 unique participants covering a balanced representation of gender and ethnicity, and includes more than 7,000 high-quality print attack sequences. Each attack is captured as a 10–20-second video with a zoom-in effect, simulating realistic presentation attack detection (PAD) scenarios and conforming to standards like NIST FATE. The images feature realistic color reproduction, no visible borders, and use flat, straight-on photo prints, enhancing consistency across samples.This dataset has been employed by certification authorities such as iBeta and NIST FATE to assess PAD systems. Its structured nature and professional collection make it an excellent resource for training and evaluating anti-spoofing models in facial recognition and biometric authentication contexts.

反照片打印攻击数据集（Anti-Spoofing Photo Print Attack Dataset）是一款专为推进面部活体检测技术研发的生物特征数据集，尤其聚焦于防范照片打印类欺骗攻击。该数据集包含超过3000名独立参与者，性别与种族分布均衡，同时收录了7000余组高质量打印攻击序列。每组攻击样本均以10至20秒的变焦视频形式采集，模拟真实的呈现攻击检测（Presentation Attack Detection, PAD）场景，且符合美国国家标准与技术研究院（National Institute of Standards and Technology, NIST）FATE等相关标准。样本图像具备逼真的色彩还原效果，无可见边框，且采用平整的正面拍摄照片打印件，有效提升了样本间的一致性。该数据集已被iBeta、NIST FATE等认证机构用于评估PAD系统。其结构化的数据集特性与专业化的采集流程，使其成为面部识别与生物特征认证场景下，训练与评估反欺骗模型的优质资源。