Silicone Mask Attack Dataset for Anti-Spoofing

The Silicone Mask Attack Dataset for Anti-Spoofing contains over 10,000 high-resolution videos simulating biometric frauds using 3D silicone masks, designed to support liveness detection and spoof-resistance research. Captured in diverse settings—multiple office spaces, residential interiors, and outdoor environments—under varied lighting (low, medium, and bright), the dataset features around 18 intricately detailed silicone masks, and actors adorned with combinations of hairstyles, glasses, wigs, and beards to enhance demographic and visual diversity. Each video incorporates natural behaviors such as head movements and blinking, making them ideal for both passive and active liveness detection model training. Recorded using a range of modern devices—including iPhone 14, iPhone 14 Pro, iPhone 13 Pro, Galaxy S23, Redmi Note 12 Pro+, Galaxy A54, Pixel 7, and Honor 70—the dataset ensures broad coverage of real-world capture scenarios. It is particularly valuable for institutions pursuing iBeta Level 2 certification, enabling developers to evaluate and harden anti-spoofing systems against sophisticated presentation attacks. The dataset is available on Kaggle (sample access) and further access may be arranged via Axon Labs.

反欺骗用硅胶面具攻击数据集（Silicone Mask Attack Dataset for Anti-Spoofing）包含超过1万条高分辨率视频，模拟使用3D硅胶面具实施的生物特征欺诈行为，旨在为活体检测与抗欺骗研究提供支撑。该数据集的采集场景覆盖多元环境，包括多间办公室、住宅室内与户外场景，光照条件涵盖弱光、中等光与强光三类；数据集包含约18款细节精细的硅胶面具，参演人员搭配不同发型、眼镜、假发与胡须造型，以增强样本在人口统计学与视觉层面的多样性。每条视频均包含头部转动、眨眼等自然动作，适用于被动式与主动式活体检测模型的训练。数据集采用多款现代设备采集，包括iPhone 14、iPhone 14 Pro、iPhone 13 Pro、Galaxy S23、Redmi Note 12 Pro+、Galaxy A54、Pixel 7与Honor 70，因此能够覆盖广泛的真实采集场景。该数据集对追求iBeta Level 2认证的机构尤为实用，可帮助开发者评估并强化反欺骗系统，以抵御复杂的呈现式攻击。该数据集可在Kaggle获取样本访问权限，完整权限可通过Axon Labs另行申请。