ArtiFact
收藏资源简介:
ArtiFact数据集由孟加拉工程技术大学创建,包含249万张图像,旨在评估合成图像检测器的泛化能力和鲁棒性。该数据集涵盖多种生成器和对象类别,模拟真实世界挑战,如社交媒体平台的图像损害。创建过程涉及从多个源随机采样真实图像,并使用25种不同的生成方法合成图像。ArtiFact数据集广泛应用于检测和防止合成图像相关的非法活动,提高图像的隐私、真实性和安全性。
The ArtiFact Dataset was developed by Bangladesh University of Engineering and Technology. It contains 2.49 million images and is designed to evaluate the generalization ability and robustness of synthetic image detectors. Covering a variety of generators and object categories, this dataset simulates real-world challenges such as image degradation on social media platforms. Its creation process involves randomly sampling real images from multiple sources and synthesizing images using 25 distinct generation methods. The ArtiFact Dataset is widely applied to detect and prevent illegal activities related to synthetic images, thereby enhancing the privacy, authenticity, and security of images.

- 1ArtiFact: A Large-Scale Dataset with Artificial and Factual Images for Generalizable and Robust Synthetic Image Detection孟加拉工程技术大学 · 2023年



