FaceForensics
收藏资源简介:
FaceForensics是一个大规模视频数据集,专注于人脸伪造检测,由慕尼黑工业大学等机构创建。该数据集包含约50万张编辑过的图像,来自超过1000个视频,用于检测和分析人脸编辑技术。数据集的创建过程涉及使用先进的面部编辑方法,自动生成重演操作,并提供源视频和目标视频的配对,以及用于分割任务的像素级二进制掩码。该数据集的应用领域包括图像取证、视频操作检测和面部重演质量改进,旨在解决数字通信中的信任问题和生物识别访问控制中的真实性验证。
FaceForensics is a large-scale video dataset dedicated to face forgery detection, developed by institutions including the Technical University of Munich and other research organizations. The dataset contains approximately 500,000 edited images derived from over 1,000 videos, which is designed for the detection and analysis of facial editing techniques. The development process of the dataset utilizes advanced facial editing methods to automatically generate reenactment operations, and provides paired source and target videos, as well as pixel-level binary masks for segmentation tasks. Application scenarios of this dataset include image forensics, video manipulation detection and facial reenactment quality improvement, aiming to address trust issues in digital communications and authenticity verification in biometric access control systems.

- 1FaceForensics: A Large-scale Video Dataset for Forgery Detection in Human Faces慕尼黑工业大学 · 2018年



