hoangnguyen311111/fraud

该数据集是一个用于训练和评估AI生成文档伪造检测模型的图像数据集。数据集包含真实文档图像和AI生成的伪造图像，共6,626张图片，其中6,326张为伪造图像（95.5%），300张为真实图像（4.5%）。数据来自16个不同的来源，包括公开的AI模型生成图像和自行收集的图像。数据集按80/20的比例分为训练集和验证集，并提供了详细的来源信息和文件命名规范。数据集主要用于计算机视觉任务，特别是文档伪造和深度伪造检测。

This dataset is an image collection designed for training and evaluating models to detect AI-generated document forgeries. It includes both real document images and AI-generated fake images, totaling 6,626 images, with 6,326 fake images (95.5%) and 300 real images (4.5%). The data comes from 16 different sources, including publicly available AI-generated images and self-collected images. The dataset is split into training and validation sets in an 80/20 ratio and provides detailed source information and file naming conventions. It is primarily used for computer vision tasks, particularly document forgery and deepfake detection.