arXiv cs.AI· Simon Freitter, David Fischinger, Silvia Poletti, Marcel Hasenbalg, Martin Boyer·· 2 天前AI 评分22
DINO-FIL：多层 Transformer 融合用于生成式图像修复伪造定位
DINO-FIL: Multi-Layer Transformer Fusion for Generative Inpainting Forgery Localization
AI 导读
DINO-FIL 是一个基于 Transformer 的框架，用于像素级定位生成式图像伪造，它结合预训练 DINOv3 ViT-H+/16 编码器的多层表示与轻量融合解码器，捕捉语义结构和编辑引入的细粒度纹理不一致。
正文 · 原文
Abstract:We propose DINO-FIL, a transformer-based framework for pixel-level localization of generative image forgeries. The method leverages multi-layer representations from a pretrained DINOv3 ViT-H+/16 encoder combined with a lightweight fusion decoder to capture both semantic structure and fine-grained texture inconsistencies introduced by image editing. By aggregating intermediate transformer features across depths, the model improves sensitivity to subtle, spatially localized artifacts that are often missed by standard single-layer approaches. We evaluate DINO-FIL in a cross-dataset setting, training on B-Free and testing on Beyond the Brush, CocoGlide, TGIF, and SAGI-D. The model achieves strong generalization, outperforming existing baselines on average and on four of five evaluated datasets, while maintaining a high true negative rate on pristine images.
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.21780 [cs.CV] |
| (or arXiv:2610.21780v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2610.21780 arXiv-issued DOI via DataCite (pending registration) |
|
| Journal reference: | ECCV Workshop AI4MFDD2026 - 2026 Workshop on AI for Multimedia Forensics & Disinformation Detection |
Submission history
From: Silvia Poletti [view email]
[v1]
Thu, 15 Oct 2026 12:36:53 UTC (34 KB)
来源：arXiv cs.AI · arxiv.org