AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

What makes fake images detectable? Understanding properties that generalize

Lucy Chai, David Bau, Ser-Nam Lim, Phillip Isola

ECCV 2020 · Image generation › Autoregressive and latent generation · approach: Generalizable fake image detection · 388 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 Xception Block 2 patch classifier instead of full Xceptionfake image detector trained on 128px PGAN · PGAN 1024pxaverage precision: 97.28 -> 99.98Improved
use Resnet Layer 1 patch classifier instead of full Resnet-18detector on SGAN CelebA-HQ · SGAN CelebA-HQaverage precision: 64.80 -> 97.22Improved
shrink receptive field too far (Xception Block 2 to Block 1) (ablation)patch classifier · PGAN 1024pxaverage precision: 99.98 -> 98.53Hurt
enlarge receptive field from Xception Block 2 to Block 5 (ablation)patch classifier · PGAN 1024pxaverage precision: 99.98 -> 91.18Hurt
add parameters without enlarging receptive field (Extended Block 2) (ablation)Xception Block 2 patch classifier · Glow, GMM, FFHQ SGANaverage precision: Glow 46.35 -> 39.23; GMM 91.38 -> 91.54; FFHQ SGAN 90.12 -> 87.82Mixed
add reprojected images to fake training set (ablation)truncated patch classifiers · Glowaverage precision: difference in AP of over 10%Hurt
add random cropping during training (ablation)Xception Block 1 patch classifier · Glow and FFHQ SGANaverage precision: Glow 82.39 -> 89.60; FFHQ SGAN 81.35 -> 79.00Mixed
save real and fake images with different formats (JPG vs PNG) (ablation)Resnet-18 detector · SGAN CelebA-HQaverage precision: 64.80 -> 100.00Improved
use different interpolation for real and fake images (ablation)Resnet-18 detector · Glowaverage precision: 47.06 -> 99.76Improved
add blur and JPEG augmentation (p=0.5)Resnet-18 detector · Glowaverage precision: 47.06 -> 97.32Improved
adversarially finetune the PGAN generator against the classifierfirst fake-image classifier · CelebA-HQ PGANaccuracy: 100% -> below 65%Hurt
use Resnet Layer 1 patch classifier instead of full Resnet-18detector trained on Face2Face · FaceForensics++ Deepfakes testaverage precision: 55.43 -> 84.39Improved

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