AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

Forensic-MoE: Exploring Comprehensive Synthetic Image Detection Traces with Mixture of Experts

Mingqi Fang, Ziguang Li, Lingyun Yu, Quanwei Yang, Hongtao Xie, Yongdong Zhang

ICCV 2025 · Image generation › Autoregressive and latent generation · approach: Generalizable fake image detection · 8 citations (OpenAlex)

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
use multiple interacting experts instead of a single universal detector (ablation)Forensic-MoE · Ablation mean over 16 test subsetsAccM: 93.78 -> 95.14Improved
add expert interaction finetuning versus independent frozen experts with AWG (ablation)Forensic-MoE · Ablation mean over 16 test subsetsAccM: 94.03 -> 95.14Improved
add EKD loss (ablation)Forensic-MoE · Ablation mean over 16 test subsetsAccM: 94.72 -> 95.14Improved
use EKD loss instead of single center loss (ablation)Forensic-MoE · Ablation mean over 16 test subsetsAccM: 94.47 -> 95.14Improved
add adaptive weight generator AWG (ablation)Forensic-MoE · Ablation mean over 16 test subsetsAccM: 94.50 -> 95.14Improved
add multi-stage adapter (ablation)Forensic-MoE expert · Ablation mean over 16 test subsetsAccM: 93.30 -> 95.14Improved
add patch decentralization (ablation)Forensic-MoE expert · Ablation mean over 16 test subsetsAccM: 90.14 -> 95.14Improved
add DWT frequency information (ablation)Forensic-MoE expert · Ablation mean over 16 test subsetsAccM: 93.79 -> 95.14Improved
train experts on GAN + diffusion data rather than 2 GANs or 2 DMs (ablation)Forensic-MoE · Ablation mean over 16 test subsetsAccImproved
increase number of experts from 1 to 2 (ablation)Forensic-MoE · Ablation mean over 16 test subsetsAcc / APImproved
add diffusion images to training of FatFormer and NPRFatFormer, NPR · ForenSynths+GenImage (trained on ProGAN+SDv1.4) overall meanAccNo clear effect
Forensic-MoE vs FatFormer on unseen generators (Kandinsky 2)Forensic-MoE · FakeInversion unseenAcc (%): 69.02 -> 88.14Improved

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