AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

Enabling Your Forensic Detector Know How Well It Performs on Distorted Samples

Bin Li, Haoyu Li, Haodong Li, Jiaming Zhong, Changsheng Chen, Jiangqun Ni, bo.cao

ICLR 2026 · Image generation › Autoregressive and latent generation · approach: Generalizable fake image detection

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add DACOM confidence routing instead of logit-based confidence routingLogit Calib. Top-1 routing · Seen distortionsmean accuracy: +5.82% in mean accuracyImproved
add DACOM confidence routing instead of logit-based confidence routingLogit Calib. Top-1 routing · Seen distortionsmean AP: +8.98% in mean APImproved
add DACOM confidence routing instead of logit-based confidence routingLogit Calib. Top-1 routing · Unseen distortionsmean accuracy: +2.62% in mean accuracyImproved
add DACOM confidence routing instead of logit-based confidence routingLogit Calib. Top-1 routing · Unseen distortionsmean AP: +8.45% in mean APImproved
use DACOM Top-1 routing instead of the single best detectorCD · Seen distortionsaverage accuracy: 84.85 -> 95.40Improved
route with DACOM instead of using worst-case detectorC2P* · Seen distortions (worst subset)accuracy: 81.10 -> 90.90Improved
add DACOM confidence filtering instead of logit confidence at 40% filteredLogit Calib. · Cross-dataset single distortionBAcc: 73.71 -> 77.32Improved
add DACOM confidence filtering instead of logit confidence at 40% filteredLogit Calib. · Cross-dataset single distortionEER: 27.50 -> 19.87Improved
choose FR-IQA target SSIM vs FSIM vs DISTS vs MS-SSIM for DACOM (ablation)DACOM · Seen distortionsaverage accuracy: 95.13 -> 95.40No clear effect
filter using no-reference IQA (ARNIQA) instead of DACOM (ablation)DACOM · Cross-dataset single distortionBAcc: 77.32 -> 69.83Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy