AxonData/replay-attack-dataset

Replay Attack Dataset是一个商业数据集，包含10,000多个重放攻击视频，涵盖基于显示器和移动设备的重放欺骗。数据集来自2,500多个独特的参与者，性别和种族分布均衡。适用于训练面部反欺骗模型和准备iBeta PAD Level 1认证。数据集结合了两种攻击流：显示重放攻击（来自各种屏幕设备，具有多样化的光照条件）和移动重放攻击（覆盖15种以上的不同手机型号）。所有视频共享相同的捕获协议：慢速相机移动，多角度覆盖。数据集特点包括：总攻击视频数超过10,000，独特的参与者超过2,500，攻击子类型包括显示重放和移动重放，移动设备覆盖15种以上的手机型号（包括Galaxy A54、Honor 70、Pixel 7），相机行为为慢速移动和多角度，性别和种族分布均衡，符合iBeta Level 1（PAD）标准。

The Replay Attack Dataset is a commercial collection of 10,000+ replay attack videos covering both display-based and mobile-based replay spoofing. Captured from 2,500+ unique participants with balanced gender and ethnicity representation, it is suitable for training face anti-spoofing models and preparing for iBeta PAD Level 1 certification. The dataset combines two attack streams: display replay attacks (from various screen devices with diverse lighting conditions) and mobile replay attacks (across 15+ different phone models). Both attack streams share the same capture protocol: slow camera movement and multi-angle coverage. Key features include: 10,000+ total attack videos, 2,500+ unique participants, attack sub-types of display replay and mobile replay, mobile device coverage of 15+ phone models (incl. Galaxy A54, Honor 70, Pixel 7), camera behavior of slow movement and multi-angle, balanced gender and ethnicity distribution, and iBeta Level 1 (PAD) compliance.