# GAP-SAM: A Global Artifact Prior for Generalizable AI-Generated Image Manipulation Localization

arXiv: https://arxiv.org/abs/2608.20929 · HTML: https://mlsanity.com/paper/2608.20929/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 18 of its 38 references that are in the map.
- First posted on arXiv 21 Aug 2026.

## Builds on (most cited first, 18 of 18)

- [SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers](https://mlsanity.com/paper/2105.15203/) (2021)
- [FiLM: Visual Reasoning with a General Conditioning Layer](https://mlsanity.com/paper/1709.07871/) (2017)
- [GenImage: A Million-Scale Benchmark for Detecting AI-Generated Image](https://mlsanity.com/paper/2306.08571/) (2023)
- [UltraEdit: Instruction-based Fine-Grained Image Editing at Scale](https://mlsanity.com/paper/2407.05282/) (2024)
- [MVSS-Net: Multi-View Multi-Scale Supervised Networks for Image Manipulation Detection](https://mlsanity.com/paper/2112.08935/) (2021)
- [SIDA: Social Media Image Deepfake Detection, Localization and Explanation with Large Multimodal Model](https://mlsanity.com/paper/2412.04292/) (2024)
- [SPAN: Spatial Pyramid Attention Network forImage Manipulation Localization](https://mlsanity.com/paper/2009.00726/) (2020)
- [Dual Data Alignment Makes AI-Generated Image Detector Easier Generalizable](https://mlsanity.com/paper/2505.14359/) (2025)
- [AutoSplice: A Text-prompt Manipulated Image Dataset for Media Forensics](https://mlsanity.com/paper/2304.06870/) (2023)
- [IML-ViT: Benchmarking Image Manipulation Localization by Vision Transformer](https://mlsanity.com/paper/2307.14863/) (2023)
- [Fake or JPEG? Revealing Common Biases in Generated Image Detection Datasets](https://mlsanity.com/paper/2403.17608/) (2024)
- [Aligned Datasets Improve Detection of Latent Diffusion-Generated Images](https://mlsanity.com/paper/2410.11835/) (2024)
- [Mesoscopic Insights: Orchestrating Multi-scale & Hybrid Architecture for Image Manipulation Localization](https://mlsanity.com/paper/2412.13753/) (2024)
- [TransForensics: Image Forgery Localization with Dense Self-Attention](https://mlsanity.com/paper/2108.03871/) (2021)
- [Rethinking Image Editing Detection in the Era of Generative AI Revolution](https://mlsanity.com/paper/2311.17953/) (2023)
- [SAFIRE: Segment Any Forged Image Region](https://mlsanity.com/paper/2412.08197/) (2024)
- [Revisiting Image Manipulation Localization under Realistic Manipulation Scenarios](https://mlsanity.com/paper/2509.20006/) (2025)
- [DiffSeg30k: A Multi-Turn Diffusion Editing Benchmark for Localized AIGC Detection](https://mlsanity.com/paper/2511.19111/) (2025)

## Abstract (arXiv)

AI-generated image manipulation localization identifies edited pixels, but its OOD performance lags behind image-level detection partly because pixel supervision entangles forensic evidence with dataset-specific mask geometry and semantic boundaries. Extending image-level distribution alignment to localization, we construct COCO-ControlNet with source-image Canny edges and depth maps to align semantics and geometry, improving OOD performance across multiple localizers. Yet tighter Mask-VAE Reconstruction Alignment (Mask-VAE) underperforms COCO-ControlNet, showing that VAE reconstruction artifacts transfer poorly to local diffusion-inpainting artifacts. We also identify boundary adhesion, where fine-tuned segmentation models snap predictions to semantic object contours rather than true manipulation boundaries. These findings motivate GAP-SAM, which encodes an image and its frozen VAE reconstruction into a global artifact token and injects it into SAM3's feature pyramid via zero-gated FiLM before pixel decoding. Without prescribing a spatial region, this token modulates dense decoding to preserve localization while suppressing semantic-boundary shortcuts. Across six datasets, GAP-SAM averages 79.8 Pixel-F1, outperforming the strongest prior method by 12.6 points. It also performs best at every tested severity of JPEG compression, Gaussian blur, and resizing.