Forgery ADE
收藏资源简介:
Forgery ADE是一个新构建的图像伪造数据集,由大连理工大学的研究团队创建。该数据集包含了使用八种当前流行的深度学习图像编辑和生成方法(四种基于GAN的方法和四种基于扩散模型的方法)对ADE 20K数据集中的图像进行部分篡改得到的图像。数据集旨在促进针对深度学习模型产生的图像伪造的通用和跨方法检测研究,共有训练集图像20,210张,测试集图像2,000张,总计177,680张图像。
Forgery ADE is a newly constructed image forgery dataset developed by the research team at Dalian University of Technology. It comprises partially tampered images generated by applying eight currently popular deep learning-based image editing and generation methods (four GAN-based methods and four diffusion model-based methods) to the images within the ADE 20K dataset. This dataset is intended to promote research on general and cross-method detection of image forgeries produced by deep learning models. It contains 20,210 training images and 2,000 test images, with a total of 177,680 images in all.




