Fujitsu-FRE/clide_synthetic_datasets

CLIDE合成图像数据集是一个由现代文本到图像模型生成的合成图像集合，按领域和生成器组织。该数据集旨在支持在领域和生成器变化下对生成图像检测方法的分析和评估。数据集包含两个视觉领域：受损汽车和发票。受损汽车图像由Dalle、Kandinsky、Midjourney、SDXL和UnCLIP等多种模型生成，而发票图像则由GPT-Image-1模型生成。数据集结构按照data/<split>/<domain>/<generator>/<image>的目录布局组织，并提供了包含文件名称、领域和生成器等信息的元数据文件。A collection of synthetic images generated by modern text-to-image models, organized by domain and generator. The dataset is designed to support analysis and evaluation of generated-image detection methods under domain and generator shifts. The dataset contains two visual domains: damaged cars and invoices. Synthetic images of damaged cars are generated by multiple text-to-image models: Dalle, Kandinsky, Midjourney, SDXL, and UnCLIP, while synthetic invoice images are generated by a single text-to-image model: GPT-Image-1. The directory layout is organized as data/<split>/<domain>/<generator>/<image>, and a metadata file is provided with information such as file name, domain, and generator.