Dissect and Prune: Enhancing Robustness in AI-Generated Image Detection

• URL: http://arxiv.org/abs/2606.10309v1
• Date: Tue, 09 Jun 2026 01:55:09 GMT
• Title: Dissect and Prune: Enhancing Robustness in AI-Generated Image Detection
• Abstract summary: existing AI-generated image detectors report high performance.<n> bias toward the real class severely limits sensitivity to generated content.<n>We propose Dissect and Prune, which leverages inpainted images to identify and prune these interfering components.
• Score: 17.159832771952136
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Abstract: While existing AI-generated image detectors report high performance, we identify that this is largely driven by a critical prediction asymmetry: a bias toward the real class that severely limits sensitivity to generated content, especially under standard post-processing operations such as compression and resizing. We hypothesize that this stems from the model's reliance on spurious features, distracting signals that obscure true generative artifacts. To address this, we propose DEAR (Dissect and Prune), which leverages inpainted images to identify and prune these interfering components. Specifically, we find that features strongly aligned to either inpainted or non-inpainted regions are less robust to post-processing. By measuring the alignment between channel activations and inpaint masks, DEAR removes features at both extremes, retaining only those that capture genuine generative artifacts. Experimental results demonstrate that our approach significantly enhances robustness against unseen generators and post-processing, effectively mitigating the prediction asymmetry. Our code is available at https://github.com/dahyedahye/dear.

Related papers

• AI-Generated Image Detectors Overrely on Global Artifacts: Evidence from Inpainting Exchange [1.2944480428047747]
  We show that VAE-based reconstruction induces a subtle but pervasive spectral shift across the entire image, including unedited regions.<n>We introduce Inpainting Exchange (INP-X), an operation that restores original pixels outside the edited region while preserving all synthesized content.<n>Our findings highlight the need for content-aware detection. Indeed, training on our dataset yields better generalization and localization than standard inpainting.
  arXiv  Detail & Related papers  (2026-01-30T09:14:10Z)
• Detecting Generated Images by Fitting Natural Image Distributions [75.31113784234877]
  We propose a novel framework that exploits geometric differences between the data manifold of natural and generated images.<n>We employ a pair of functions engineered to yield consistent outputs for natural images but divergent outputs for generated ones.<n>An image is identified as generated if a transformation along its data manifold induces a significant change in the loss value of a self-supervised model pre-trained on natural images.
  arXiv  Detail & Related papers  (2025-11-03T07:20:38Z)
• Active Adversarial Noise Suppression for Image Forgery Localization [56.98050814363447]
  We introduce an Adversarial Noise Suppression Module (ANSM) that generate a defensive perturbation to suppress the attack effect of adversarial noise.<n>To our best knowledge, this is the first report of adversarial defense in image forgery localization tasks.
  arXiv  Detail & Related papers  (2025-06-15T14:53:27Z)
• Stay-Positive: A Case for Ignoring Real Image Features in Fake Image Detection [31.23513370504603]
  We argue that an image should be classified as fake if and only if it contains artifacts introduced by the generative model.<n>We propose Stay Positive, an algorithm designed to constrain the detectors focus to generative artifacts while disregarding those associated with real data.
  arXiv  Detail & Related papers  (2025-02-11T18:59:10Z)
• Understanding and Improving Training-Free AI-Generated Image Detections with Vision Foundation Models [68.90917438865078]
  Deepfake techniques for facial synthesis and editing pose serious risks for generative models.<n>In this paper, we investigate how detection performance varies across model backbones, types, and datasets.<n>We introduce Contrastive Blur, which enhances performance on facial images, and MINDER, which addresses noise type bias, balancing performance across domains.
  arXiv  Detail & Related papers  (2024-11-28T13:04:45Z)
• Improving Synthetic Image Detection Towards Generalization: An Image Transformation Perspective [45.210030086193775]
  Current synthetic image detection (SID) pipelines are primarily dedicated to crafting universal artifact features.<n>We propose SAFE, a lightweight and effective detector with three simple image transformations.<n>Our pipeline achieves a new state-of-the-art performance, with remarkable improvements of 4.5% in accuracy and 2.9% in average precision against existing methods.
  arXiv  Detail & Related papers  (2024-08-13T09:01:12Z)
• Improving Accuracy-robustness Trade-off via Pixel Reweighted Adversarial Training [61.61368146268329]
  We propose Pixel-reweighted AdveRsarial Training (PART), a new framework that partially reduces $epsilon$ for less influential pixels. Part achieves a notable improvement in accuracy without compromising robustness on CIFAR-10, SVHN and TinyImagenet-200.
  arXiv  Detail & Related papers  (2024-06-02T09:43:34Z)
• Beyond the Spectrum: Detecting Deepfakes via Re-Synthesis [69.09526348527203]
  Deep generative models have led to highly realistic media, known as deepfakes, that are commonly indistinguishable from real to human eyes. We propose a novel fake detection that is designed to re-synthesize testing images and extract visual cues for detection. We demonstrate the improved effectiveness, cross-GAN generalization, and robustness against perturbations of our approach in a variety of detection scenarios.
  arXiv  Detail & Related papers  (2021-05-29T21:22:24Z)
• High-Resolution Image Inpainting with Iterative Confidence Feedback and Guided Upsampling [122.06593036862611]
  Existing image inpainting methods often produce artifacts when dealing with large holes in real applications. We propose an iterative inpainting method with a feedback mechanism. Experiments show that our method significantly outperforms existing methods in both quantitative and qualitative evaluations.
  arXiv  Detail & Related papers  (2020-05-24T13:23:45Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.