Do Transformations Reveal the Truth? Generative Residual Learning for Generalized AI-Generated Image Detection

2026-07-09

AI Summary

This work addresses the limited generalization capability of existing methods in detecting generative AI–produced images, particularly when confronted with diverse or previously unseen generation models. To tackle this challenge, the authors propose the GenRes and GenRes++ frameworks, which uniquely integrate neural tensor networks with a learnable attention mechanism to model fine-grained relationships between original images and their multi-transformed variants. A key innovation is the introduction of the PE-Core feature extractor, which significantly enhances cross-domain generalization. Extensive experiments demonstrate that the proposed approach consistently outperforms current state-of-the-art methods across multiple benchmark datasets, exhibiting superior detection performance and robustness against unseen generative models.

Abstract

The rapid advancement of generative AI has enabled the creation of highly realistic deepfake media, posing significant threats, including misinformation, digital identity theft, fraud, and manipulation of public opinion. AI-generated image (AIGI) detection is reliably challenging due to the diversity of generative methods and the subtle artifacts they leave behind. In this work, we propose GenRes, a novel framework for generative residual learning via a neural tensor network, which models fine-grained relational features between original and transformed samples to enhance generalization. To address scenarios involving multiple generative transformations, we introduce GenRes++, which employs a learnable attention mechanism to aggregate relational features across multiple transformed samples and enables the model to focus on the most informative cues. Both models leverage PE-Core as a feature extractor, providing generalized and semantically rich embeddings that improve cross-domain performance and enable the detection of AIGI generated by unseen methods. Comprehensive experiments on multiple benchmark datasets demonstrate that the proposed GenRes++ approach outperforms existing methods.

Problem

• AI-generated image detection
• deepfake
• generalization
• generative AI
• image forensics

Innovation

• Generative Residual Learning
• Neural Tensor Network
• Learnable Attention Mechanism
• Cross-domain Generalization
• AI-Generated Image Detection