ML Security PapersRealHD: A High-Quality Dataset for Robust Detection of State-of-the-Art AI-Generated Images
Hanzhe Yu 1, Yun Ye 2, Jintao Rong 1, Qi Xuan 1,3, Chen Ma 1,3
Published on arXiv
2602.10546
Output Integrity Attack
OWASP ML Top 10 — ML09
Key Finding
Detection models trained on RealHD achieve superior generalization over models trained on existing datasets, with the NLM noise-entropy baseline delivering competitive performance across diverse generation methods
RealHD / NLM Noise Entropy Detection
Novel technique introduced
The rapid advancement of generative AI has raised concerns about the authenticity of digital images, as highly realistic fake images can now be generated at low cost, potentially increasing societal risks. In response, several datasets have been established to train detection models aimed at distinguishing AI-generated images from real ones. However, existing datasets suffer from limited generalization, low image quality, overly simple prompts, and insufficient image diversity. To address these limitations, we propose a high-quality, large-scale dataset comprising over 730,000 images across multiple categories, including both real and AI-generated images. The generated images are synthesized via state-of-the-art methods, including text-to-image generation (guided by over 10,000 carefully designed prompts), image inpainting, image refinement, and face swapping. Each generated image is annotated with its generation method and category. Inpainting images further include binary masks to indicate inpainted regions, providing rich metadata for analysis. Compared to existing datasets, detection models trained on our dataset demonstrate superior generalization capabilities. Our dataset not only serves as a strong benchmark for evaluating detection methods but also contributes to advancing the robustness of AI-generated image detection techniques. Building upon this, we propose a lightweight detection method based on image noise entropy, which transforms the original image into an entropy tensor of Non-Local Means (NLM) noise before classification. Extensive experiments demonstrate that models trained on our dataset achieve strong generalization, and our method delivers competitive performance, establishing a solid baseline for future research. The dataset and source code are publicly available at https://real-hd.github.io.
Key Contributions
- RealHD dataset: 730K+ high-quality images from text-to-image generation, inpainting, refinement, and face swapping, with rich annotations including generation method, category, and inpainting masks
- Lightweight AI-generated image detector based on image noise entropy using Non-Local Means (NLM) noise transformed into an entropy tensor for classification
- Demonstrated superior generalization of detection models trained on RealHD compared to existing datasets
🛡️ Threat Analysis
The paper's primary contribution is a benchmark dataset and detection method for identifying AI-generated images (deepfakes, synthetic imagery from diffusion models, face swapping, inpainting) — directly addressing AI-generated content detection and output authenticity, which is the core of ML09.