ImageAttributionBench: How Far Are We from Generalizable Attribution?

2026-05-13

AI Summary

Existing image provenance datasets suffer from limited scale, outdated generation methods, and insufficient semantic diversity, hindering the development of robust and generalizable attribution models. To address this gap, this work introduces ImageAttributionBench, the first large-scale benchmark systematically constructed to encompass a wide array of state-of-the-art generative models and real-world semantic domains. The benchmark features challenging evaluation protocols, including semantically disjoint splits and image degradation scenarios, designed to better reflect practical conditions. Comprehensive evaluations reveal that current leading methods exhibit significant performance degradation under novel semantic content or degraded inputs, exposing fundamental limitations in their generalization capabilities. This study thus establishes a more realistic and rigorous benchmark for image provenance research and delineates clear directions for future improvements.

Abstract

The rapid advancement of generative AI has enabled the creation of highly realistic and diverse synthetic images, posing critical challenges for image provenance and misinformation detection. This underscores the urgent need for effective image attribution. However, existing attribution datasets are constrained by limited scale, outdated generation methods, and insufficient semantic diversity - hindering the development of robust and generalizable attribution models. To address these limitations, we introduce ImageAttributionBench, a comprehensive dataset comprising images synthesized by a wide array of advanced generative models with state-of-the-art (SOTA) architectures. Covering multiple real-world semantic domains, the dataset offers rich diversity and scale to support and accelerate progress in image attribution research. To simulate real-world attribution scenarios, we evaluate several SOTA attribution methods on ImageAttributionBench under two challenging settings: (1) training on a standard balanced split and testing on degraded images, and (2) training and testing on semantically disjoint splits. In both cases, current methods exhibit consistently poor performance, revealing significant limitations in their robustness and generalization to unseen semantic content. Our work provides a rigorous benchmark to facilitate the development and evaluation of future image attribution methods.

Problem

• image attribution
• generative AI
• synthetic images
• generalization
• misinformation detection

Innovation

• image attribution
• generative AI
• benchmark dataset
• generalization
• synthetic image detection