AI Paper Atlas

Research map / Image generation / GANs

GIQA: Generated Image Quality Assessment

Shuyang Gu, Jianmin Bao, Dong Chen, Fang Wen

ECCV 2020 · Image generation › GANs · approach: GAN latent optimization and evaluation · 92 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 binary classifiers instead of direct regression of pseudo label (ablation)MBC-GIQA · LGIQA-FFHQpairwise accuracy: 0.678 -> 0.731Improved
use multiple binary classifiers instead of a single real/fake classifier (ablation)MBC-GIQA · LGIQA-FFHQpairwise accuracy: 0.663 -> 0.731Improved
use Gaussian mixture instead of a single Gaussian (ablation)GMM-GIQA · LGIQA-FFHQpairwise accuracy: 0.620 -> 0.764Improved
use GMM-GIQA rather than SGM-GIQA on LSUN-cat (ablation)GMM-GIQA · LGIQA-LSUN-catpairwise accuracy: 0.829 -> 0.846Improved
add OHEM weighting on low-quality samples in discriminator (ablation)StyleGAN · FFHQ (256x256)FID: 17.35 -> 16.89Improved
add OHEM weighting on low-quality samples in discriminator (ablation)StyleGAN · LSUN-cat (256x256)FID: 18.67 -> 18.12Improved
add image picker to discard low-quality samples (ablation)StyleGAN+OHEM · LSUN-cat (256x256)FID: 18.12 -> 16.25Improved
add image picker to discard low-quality samples (ablation)StyleGAN+OHEM · FFHQ (256x256)FID: 16.89 -> 16.68Improved
use image picker at remaining rate 0.9 versus truncation rate 0.9StyleGAN · LSUN-catFID: 18.05 -> 16.65Improved
increase number of binary classifiers beyond 8 (ablation)MBC-GIQA · LGIQA-FFHQpairwise accuracy: 0.731 -> 0.717Hurt
use training patch size 192 instead of 256 (ablation)MBC-GIQA · LGIQA-FFHQpairwise accuracy: 0.690 -> 0.731Improved
increase Gaussian components from 70 to 100 (ablation)GMM-GIQA · LGIQA-FFHQpairwise accuracy: 0.764 -> 0.753Hurt

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