high_precision_3d_resin_masks_real_faces
收藏资源简介:
Rubber Mask Attack Dataset是一个专业的面部反欺骗数据集,针对面部识别系统和生物识别认证,采用定制的橡胶面具攻击。该数据集采用双录制方法,捕捉同一目标个体的真实面部视频和相应的橡胶面具攻击视频。数据集特点包括约2k个视频,涵盖6个身份ID,包含原始面孔视频,具有活性的生活特征,如缩放、旋转、头部倾斜等,以及不同的发型和配饰,室内不同光照条件的多样性。主要用于iBeta Level 2认证的准备,以及用于欺骗系统的主动和被动生活检测。
Rubber Mask Attack Dataset is a professional facial anti-spoofing dataset designed for facial recognition systems and biometric authentication applications, employing custom rubber mask spoofing attacks. This dataset adopts a dual-recording approach to capture both genuine facial videos and corresponding rubber mask attack videos of the same target individual. The dataset features approximately 2,000 videos covering 6 unique identity IDs, including raw facial videos that exhibit realistic liveness-related dynamics such as zooming, rotation, and head tilting, alongside various hairstyles, accessories, and diverse indoor lighting conditions. It is primarily intended for the preparation of iBeta Level 2 certification, as well as for active and passive liveness detection against spoofing attacks.
高精度3D树脂面具与真实人脸反欺诈数据集
数据集概述
- 许可证类型: CC-BY-NC-4.0
- 任务类别: 图像特征提取、图像分类
- 语言: 英语
- 标签: 金融、医疗
核心特征
- 数据规模: 约2000个视频,拍摄自6个不同ID
- 内容构成: 包含原始人脸视频及对应树脂面具攻击视频
- 活体检测特征: 缩放(放大/缩小)、左右旋转、头部倾斜等动作
- 多样性特征: 不同发型、配饰及室内多种光照条件
- 主要应用: iBeta Level 2认证准备、主动/被动活体检测系统开发
商业版本
- 完整商业版本需通过Axonlabs官网申请购买
应用场景
- iBeta Level 2认证合规
- 模型训练与预认证测试
- 内部活体检测模型开发
- 反欺诈模型训练与验证
- 算法漏洞检测
相关数据集




