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

TL;DR

DRAGON is a large, diverse dataset of images generated by 25 diffusion models, aimed at improving detection of synthetic images and supporting forensic research.

Contribution

It introduces a comprehensive, multi-size dataset of diffusion-generated images and a prompt expansion pipeline to enhance image realism for detection tasks.

Findings

Enhanced image realism through prompt expansion improves quality metrics.

Dataset covers both recent and older diffusion models for broad applicability.

Provides a benchmark test set for evaluating detection methods.

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to the rapid evolution of the field, existing datasets often cover only a limited range of models and quickly become outdated.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.