AI Paper Atlas

Research map / Image generation / Super-resolution

RankSRGAN: Generative Adversarial Networks With Ranker for Image Super-Resolution

Wenlong Zhang, Yihao Liu, Chao Dong, Yu Qiao

ICCV 2019 · Image generation › Super-resolution · approach: CNN and transformer SR architectures · 421 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 a Ranker with rank-content loss to SRGANSRGAN · PIRM-TestNIQE: 2.71 -> 2.51Improved
add a Ranker with rank-content loss to SRGANSRGAN · BSD100NIQE: 3.29 -> 3.01Improved
add a Ranker with rank-content loss to ESRGAN-level baseline comparisonESRGAN · Set14NIQE: 3.28 -> 3.28No clear effect
train Ranker with ground-truth HR as third algorithm instead of ESRGAN (ablation)RankSRGAN · PIRM-TestPSNR: 25.60 -> 26.00Improved
train Ranker with ground-truth HR as third algorithm instead of ESRGAN (ablation)RankSRGAN · PIRM-TestNIQE: 2.51 -> 2.58Hurt
use Ma-based Ranker instead of NIQE-based Ranker (ablation)RankSRGAN · PIRM-TestNIQE: 2.51 -> 2.65Hurt
use Ma-based Ranker instead of NIQE-based Ranker (ablation)RankSRGAN · PIRM-Test10-Ma: 1.39 -> 1.38Improved
use PI-based Ranker instead of NIQE-based Ranker (ablation)RankSRGAN · PIRM-TestPI: 1.95 -> 1.94Improved
use margin-ranking Ranker instead of regression Ranker (NIQE labels) (ablation)RankSRGAN · PIRM-TestNIQE: 2.53 -> 2.51Improved
use margin-ranking Ranker instead of regression Ranker (Ma labels) (ablation)RankSRGAN · PIRM-Test10-Ma: 1.43 -> 1.38Improved
use margin-ranking Ranker instead of regression Ranker (Ma labels) (ablation)RankSRGAN · PIRM-TestNIQE: 2.61 -> 2.65Hurt
label rank dataset by metric rank instead of model classification (ablation)RankSRGAN · PIRM-TestNIQEImproved

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