AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

Detecting AI-Generated Forgeries via Iterative Manifold Deviation Amplification

Jiangling Zhang, Shuxuan Gao, Bofan Liu, Siqiang Feng, Jirui Huang, Yaxiong Chen, Ziyu Chen

CVPR 2026 · Image generation › Autoregressive and latent generation · approach: Generalizable fake image detection · 2 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
add dual-stream segmentation network (DSSN) (ablation)single-stream baseline · GIT averageIoU: 0.670 -> 0.684Improved
add DSSN (ablation)single-stream baseline · TT averageIoU: 0.459 -> 0.483Improved
add TAPI module on top of DSSN (ablation)IFA-Net · GIT averageIoU: 0.684 -> 0.745Improved
add TAPI module on top of DSSN (ablation)IFA-Net · TT averageF1: 0.625 -> 0.670Improved
add Adaptive Decoder (ablation)IFA-Net · GIT averageF1: 0.823 -> 0.855Improved
add Adaptive Decoder (ablation)IFA-Net · TT averageIoU: 0.554 -> 0.586Improved
full IFA-Net vs single-stream baseline (ablation)IFA-Net · GIT averageIoU: 0.670 -> 0.778Improved
IFA-Net vs strongest baseline DcDsDiff on traditional tamperingIFA-Net · TT averageIoU: 0.625 -> 0.586Hurt
IFA-Net vs DcDsDiff on diffusion tamperingIFA-Net · GIT averageIoU: 0.713 -> 0.778Improved
IFA-Net vs DcDsDiff under strong Gaussian blur and heavy JPEGIFA-Net · GIT/TT robustnessF1Mixed

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