Do Transformations Reveal the Truth? Generative Residual Learning for Generalized AI-Generated Image Detection
2026-07-09 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors explore how to detect AI-generated images, which can be very realistic and hard to spot. They create a system called GenRes that compares subtle differences between original and altered images using a special neural network. They improve this to GenRes++ by adding an attention mechanism that helps focus on the most important features when images go through multiple changes. Their methods use a strong feature extractor to work well even with new types of AI-generated images. Tests show their approach detects fake images better than existing tools.
generative AIdeepfake detectionneural tensor networkattention mechanismfeature extractioncross-domain performanceimage transformationrelational featuresAIGI detectionbenchmark datasets
Authors
Kutub Uddin, Nusrat Tasnim, Awais Khan, Mohammad Umar Farooq, Khalid Malik
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.