iBeta Level 2 Dataset

The iBeta Level 2 Dataset is a publicly available dataset hosted on Kaggle, designed to support research and development in presentation attack detection (PAD) systems. It contains a collection of biometric samples, primarily focused on facial and iris recognition, with an emphasis on detecting spoofing attempts using various artificial means. The dataset includes both genuine and fraudulent samples, such as printed photos, digital screens, and 3D masks, to simulate real-world attack scenarios. With over several thousand high-resolution images and corresponding metadata, this dataset enables the training and validation of machine learning models for enhanced security applications. It is particularly useful for developing algorithms that can distinguish between live subjects and presentation attacks, thereby improving the reliability of biometric authentication systems. Researchers and developers can utilize this resource to advance anti-spoofing techniques and contribute to more secure identity verification processes.

iBeta Level 2数据集（iBeta Level 2 Dataset）是托管于Kaggle平台的公开可用数据集，旨在为呈现攻击检测（presentation attack detection, PAD）系统的研发提供支撑。该数据集收录了一系列生物特征样本，核心聚焦于面部与虹膜识别领域，重点针对各类人工伪造手段下的欺骗攻击检测任务。数据集涵盖真实样本与欺诈样本两大类别，其中欺诈样本包含打印照片、数字屏幕、3D面具等，用以模拟真实场景中的攻击案例。该数据集包含数千张高分辨率图像及其对应元数据，可用于训练和验证面向增强型安全应用的机器学习模型，尤其适用于开发可区分活体样本与呈现攻击的算法，从而提升生物特征认证系统的可靠性。研究人员与开发者可依托该资源推进反欺骗技术的研发，助力构建更安全的身份验证流程。