AEDR: Training-Free AI-Generated Image Attribution via Autoencoder Double-Reconstruction

2025-07-25

AI Summary

Attribution of AI-generated images remains challenging, particularly for advanced diffusion models, where existing reconstruction-based attribution methods suffer from low accuracy and high computational overhead. Method: This paper proposes a training-free dual-reconstruction attribution framework that applies the same autoencoder to reconstruct an input image twice; the ratio of the two reconstruction losses serves as a robust attribution signal, while an image homogeneity metric is introduced to calibrate complexity-induced bias. Contribution/Results: This work pioneers the integration of dual-reconstruction mechanisms with loss-ratio analysis, achieving both high accuracy and efficiency. Extensive experiments across eight state-of-the-art diffusion models demonstrate a 25.5% improvement in attribution accuracy over prior methods, with computational latency reduced to merely 1% of the current SOTA approach.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Adversarial Learning & Robustness
• Humans and AI: Other Foundations of Human Computation & AI

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

The rapid advancement of image-generation technologies has made it possible for anyone to create photorealistic images using generative models, raising significant security concerns. To mitigate malicious use, tracing the origin of such images is essential. Reconstruction-based attribution methods offer a promising solution, but they often suffer from reduced accuracy and high computational costs when applied to state-of-the-art (SOTA) models. To address these challenges, we propose AEDR (AutoEncoder Double-Reconstruction), a novel training-free attribution method designed for generative models with continuous autoencoders. Unlike existing reconstruction-based approaches that rely on the value of a single reconstruction loss, AEDR performs two consecutive reconstructions using the model's autoencoder, and adopts the ratio of these two reconstruction losses as the attribution signal. This signal is further calibrated using the image homogeneity metric to improve accuracy, which inherently cancels out absolute biases caused by image complexity, with autoencoder-based reconstruction ensuring superior computational efficiency. Experiments on eight top latent diffusion models show that AEDR achieves 25.5% higher attribution accuracy than existing reconstruction-based methods, while requiring only 1% of the computational time.

Problem

• Tracing origin of AI-generated images for security
• Improving accuracy in reconstruction-based attribution methods
• Reducing computational costs in image attribution

Innovation

• Training-free AI-generated image attribution method
• Double-reconstruction with autoencoder loss ratio
• Calibration via image homogeneity for accuracy