AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

Beyond Semantic Features: Pixel-level Mapping for Generalized AI-Generated Image Detection

Chenming Zhou, Jiaan Wang, Yu Li, Lei Li, Juan Cao, Sheng Tang

AAAI 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 fixed pixel-level mapping preprocessingResNet-50 · GenImagemAcc: 67.0 -> 98.4Improved
add fixed pixel-level mapping preprocessingResNet-50 (ProGAN-trained) · Self-Synthesis 9 GANsmean accuracy: 93.2 -> 97.9 vs NPRImproved
use random per-sample mapping instead of fixed mapping (ablation)pixel-level mapping · GenImagemAcc: 98.4 -> 97.5Hurt
use random per-sample mapping instead of fixed mapping (ablation)pixel-level mapping · Self-Synthesis 9 GANsmean accuracy: 97.9 -> 96.9Hurt
apply high-pass filtering before ResNet-50 (ablation)ResNet-50 · GenImagemAcc: 67.0 -> 64.4Hurt
apply patch shuffling size 8 before ResNet-50 (ablation)ResNet-50 · GenImagemAcc: 67.0 -> 70.7Improved
apply patch shuffling size 2 before ResNet-50 (ablation)ResNet-50 · GenImagemAcc: 67.0 -> 50.5Hurt

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