DRAGON: A Large-Scale Dataset of Realistic Images Generated by Diffusion Models

2025-05-16

AI Summary

The proliferation of diffusion-model-generated images in disinformation campaigns has outpaced existing detection methods, which rely on outdated and narrowly scoped training data. Method: This paper introduces DRAGON, a large-scale synthetic image dataset covering 25 mainstream and state-of-the-art diffusion models. It innovatively integrates LLM-enhanced prompt engineering with multi-scale image generation to substantially improve visual diversity and photorealism. Image quality is rigorously validated using standard metrics—including CLIP Score and FID—demonstrating superior performance over existing benchmarks. Contribution/Results: DRAGON comprises high-fidelity samples spanning diverse semantic themes and model evolution stages, along with a dedicated test set. It establishes the first systematic, extensible, and authoritative benchmark for AI-generated content detection and provenance analysis, enabling reproducible, up-to-date evaluation of forensic methods.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Generation

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Web Mining and Content Analysis: Large pretrained models with web data
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

The remarkable ease of use of diffusion models for image generation has led to a proliferation of synthetic content online. While these models are often employed for legitimate purposes, they are also used to generate fake images that support misinformation and hate speech. Consequently, it is crucial to develop robust tools capable of detecting whether an image has been generated by such models. Many current detection methods, however, require large volumes of sample images for training. Unfortunately, due to the rapid evolution of the field, existing datasets often cover only a limited range of models and quickly become outdated. In this work, we introduce DRAGON, a comprehensive dataset comprising images from 25 diffusion models, spanning both recent advancements and older, well-established architectures. The dataset contains a broad variety of images representing diverse subjects. To enhance image realism, we propose a simple yet effective pipeline that leverages a large language model to expand input prompts, thereby generating more diverse and higher-quality outputs, as evidenced by improvements in standard quality metrics. The dataset is provided in multiple sizes (ranging from extra-small to extra-large) to accomodate different research scenarios. DRAGON is designed to support the forensic community in developing and evaluating detection and attribution techniques for synthetic content. Additionally, the dataset is accompanied by a dedicated test set, intended to serve as a benchmark for assessing the performance of newly developed methods.

Problem

• Detect fake images generated by diffusion models
• Address outdated datasets with limited model coverage
• Enhance realism and diversity in synthetic image detection

Innovation

• Comprehensive dataset from 25 diffusion models
• LLM-enhanced prompts for realistic image generation
• Multiple dataset sizes for varied research needs