Metraipolis › archive › dc_5efd13d4526b
EIB-Net: Entropy-Guided Information Bottleneck for Generalizable AI-Generated Image Detection — arXiv:2609.29064
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.29064
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-08T04:27:15.932Z
- SHA-256 of the archived text
c8beb76b592355dadc1d54570d47dba8a12d00eadf03c64913aef7043a3a34f5- Stable URL of this page
https://metraipolis.com/documento/dc_5efd13d4526b
How to cite
EIB-Net: Entropy-Guided Information Bottleneck for Generalizable AI-Generated Image Detection — arXiv:2609.29064 — arXiv (en), https://arxiv.org/abs/2609.29064 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_5efd13d4526b (Metraipolis archive dc_5efd13d4526b, sha256 c8beb76b592355dadc1d54570d47dba8a12d00eadf03c64913aef7043a3a34f5)
BibTeX
@misc{dc5efd13d4526b,
title = {EIB-Net: Entropy-Guided Information Bottleneck for Generalizable AI-Generated Image Detection — arXiv:2609.29064},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.29064},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_5efd13d4526b; sha256 c8beb76b592355dadc1d54570d47dba8a12d00eadf03c64913aef7043a3a34f5},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.29064. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
The proliferation of photorealistic AI-generated images demands robust detection methods that generalize across diverse generative models. While existing approaches target manipulation-based forgeries with local artifacts, generation-based images (e.g., from diffusion models) lack such traces, posing a fundamental challenge. We observe that generative models prioritize global semantics at the expense of local texture fidelity, making low-texture regions key indicators of synthetic origin. To exploit this, we propose EIB-Net, an Entropy-guided Information Bottleneck Network. EIB-Net introduces a novel Image Entropy (IE) metric to automatically select the most informative (lowest-entropy) patch, then processes it with a Variational Information Bottleneck (VIB) to learn compact, generalizable features. Extensive experiments on DIFF, DiffusionForensics, and GenImage benchmarks demonstrate state-of-the-art performance: EIB-Net achieves 85.7\% accuracy using only 2\% of training data, outperforming full-image baselines by over 15\%, and maintains robust cross-generator generalization (83.5\% average accuracy on GenImage). Furthermore, our entropy-guided patch selection (EGPL) consistently enhances diverse backbones (CNNs and Transformers), proving its practical value for data-efficient detection. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-25 in cs.AI). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.29064 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-08T03:50:57.999Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-25?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).