CASIA-FASD
收藏资源简介:
CASIA-FASD数据集是由中国科学院自动化研究所创建的,主要用于人脸反欺骗(Face Anti-Spoofing)任务。该数据集包含三种类型的假脸攻击,包括扭曲照片攻击、剪切照片攻击和视频攻击,共有56个图像标注对。数据集的创建过程结合了专家知识,对四种常见的欺骗证据进行了详细标注。CASIA-FASD数据集主要用于评估可解释人脸反欺骗方法的解释质量,旨在提高人脸识别系统的透明度和用户信任度。
CASIA-FASD dataset was developed by the Institute of Automation, Chinese Academy of Sciences, and is primarily used for the Face Anti-Spoofing task. This dataset includes three types of fake face attacks, namely distorted photo attacks, cropped photo attacks and video attacks, with a total of 56 image annotation pairs. During its construction, expert knowledge was incorporated, and detailed annotations were made for four common spoofing cues. The CASIA-FASD dataset is mainly utilized to evaluate the explanatory quality of interpretable face anti-spoofing methods, with the goal of improving the transparency and user trust of face recognition systems.




