CASIA-SURF

CASIA-SURF是由中国科学院自动化研究所创建的大型多模态人脸反欺骗数据集，包含1000个主题的21000个视频，涵盖RGB、深度和红外三种模态。数据集通过多种攻击手段（如打印和切割打印攻击）进行丰富，旨在提高人脸识别系统的安全性。创建过程中，研究者们采用了先进的传感器技术和深度学习方法，确保数据集的质量和多样性。该数据集广泛应用于多模态人脸反欺骗技术的研究和开发，特别是在提高模型对各种模态和攻击类型的适应性方面。

CASIA-SURF is a large-scale multimodal face anti-spoofing dataset created by the Institute of Automation, Chinese Academy of Sciences. It consists of 21,000 videos from 1,000 subjects, covering three modalities: RGB, depth, and infrared. The dataset is enriched with various attack types such as printed and cut-print attacks, aiming to enhance the security of facial recognition systems. During its development, researchers adopted advanced sensor technologies and deep learning methods to ensure the quality and diversity of the dataset. This dataset has been widely applied in the research and development of multimodal face anti-spoofing technologies, particularly in improving the model's adaptability to diverse modalities and attack categories.