Simple Domain Generalization for Strong Pixel-Level Image Tampering Detection in Modern VLMs
2026-07-20 • Computer Vision and Pattern Recognition
Computer Vision and Pattern RecognitionArtificial Intelligence
AI summaryⓘ
The authors studied how to detect image tampering caused by different advanced vision-language models (VLMs), which can be tricky especially when the tampering comes from unfamiliar models. They created a new training approach that balances the number of real and tampered images in each training step to avoid bias, and also gradually adapts the detector using small amounts of new data without overfitting. Their method improved detection accuracy significantly compared to previous work across several modern VLMs. This shows their simple strategies help the tampering detector stay effective even when facing new, unseen image manipulations.
Vision-Language ModelsImage Tampering DetectionDomain GeneralizationPixel-level LocalizationOut-of-DistributionBalanced Minibatch SamplingLate-injection TraininggIoUcIoU
Authors
Yi Tang, Xinyi Shang, Jiacheng Cui, Sondos Mahmoud Bsharat, Jiacheng Liu, Xiaohan Zhao, Tran Dinh Tien, Ahmed Elhagry, Salwa K. Al Khatib, Tianjun Yao, Yonina C. Eldar, Jing-Hao Xue, Hao Li, Salman Khan, Zhiqiang Shen
Abstract
Modern vision-language models (VLMs) have significantly improved image generation and editing capabilities, making pixel-level image tampering detection increasingly important yet challenging under cross-model and out-of-distribution shifts. This work studies domain generalization for pixel-level image tampering detection in modern VLMs like ChatGPT, Gemini, Qwen-Image, etc., aiming to learn tampering localization models that remain robust across diverse VLM-generated manipulation distributions. We propose a simple yet effective domain-generalized training framework built on two practical strategies. First, we introduce a balanced minibatch sampling scheme that strategically samples tampered and real images in each minibatch, preventing biased optimization toward either manipulated artifacts or clean-image priors and avoiding training collapse, ensuring that each optimization step receives proper sampled gradient signals. Second, we adopt a simple late-injection strategy, where the detector is first trained on large-scale base data until stable convergence, and then exposed to a small amount of newly selected supporting data from emerging VLM distributions, improving adaptability without overfitting to limited new domains. Together, these components provide a simple yet strong recipe for improving pixel-level tampering localization and OOD robustness across modern VLMs. Despite the conceptual simplicity, our framework outperforms the prior state-of-the-art PIXAR by a large margin of 26.1% and 26.8% relative improvement in average gIoU and cIoU, respectively, across OOD VLMs of GPT-Images-2.0, Gemini-3.1, FLUX.2, and Seedream 4.5. Our code is available at https://github.com/VILA-Lab/PIXAR-DG