Research map / Image generation / Super-resolution
RankSRGAN: Super Resolution Generative Adversarial Networks With Learning to Rank
Wenlong Zhang, Yihao Liu, Chao Dong, Yu Qiao
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add rank-content loss from a Ranker to SRGAN | SRGAN · PIRM-Test x4 | NIQE: 2.71 -> 2.51 | Improved |
| add rank-content loss from a Ranker to SRGAN | SRGAN · BSD100 x4 | NIQE: 3.29 -> 3.01 | Improved |
| add rank-content loss from a Ranker to SRGAN | SRGAN · PIRM-Test x4 | PSNR: 25.60 -> 25.62 | No clear effect |
| label rank pairs by metric rank instead of model classification (ablation) | RankSRGAN · PIRM-Test x4 | NIQE | Improved |
| use ground-truth HR instead of ESRGAN as best rank-dataset level (ablation) | RankSRGAN · PIRM-Test x4 | PSNR: NIQE 2.51 -> 2.58, PSNR 25.60 -> 26.00 | Mixed |
| train Ranker with margin-ranking instead of MSE regression (ablation) | RankSRGAN (NIQE Ranker) · PIRM-Test x4 | NIQE: 2.53 -> 2.51 | Improved |
| train Ranker with margin-ranking instead of MSE regression (ablation) | RankSRGAN (Ma Ranker) · PIRM-Test x4 | 10-Ma: 1.43 -> 1.38 | Improved |
| rank labels from PI instead of NIQE (ablation) | RankSRGAN · PIRM-Test x4 | PI: 1.95 -> 1.94 | Improved |
| add MSE loss weight alpha2 to rank-content loss (ablation) | RankSRGAN · PIRM-Test x4 | PSNR: PSNR 25.62 -> 26.62, NIQE 2.51 -> 2.72 | Mixed |
| use deeper Ranker VGG-12 instead of VGG-8 (ablation) | Ranker · Rank validation set | SROCC: 0.83 -> 0.88 | Improved |
| use VGG-16 instead of VGG-12 Ranker (ablation) | Ranker · Rank validation set | SROCC: 0.88 -> 0.88 | No clear effect |
| change monotonic function (sigmoid, exponential, identity) of rank-content loss (ablation) | RankSRGAN · PIRM-Test x4 | NIQE | No clear effect |
Later papers that compared with this method
4 later accepted papers report RankSRGAN as a baseline in their results tables; 3 of them report a better number on most of the results they share with it.
- Deterministic Image-to-Image Translation via Denoising Brownian Bridge Models with Dual Approximators (CVPR 2025 · better on 1 of 1 shared result)
- Best-Buddy GANs for Highly Detailed Image Super-resolution (AAAI 2022 · better on 2 of 3 shared results)
- Debiased Subjective Assessment of Real-World Image Enhancement (CVPR 2021 · better on 0 of 1 shared result)
- Hierarchical Conditional Flow: A Unified Framework for Image Super-Resolution and Image Rescaling (ICCV 2021 · better on 1 of 1 shared result)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- RankSRGAN: Generative Adversarial Networks With Ranker for Image Super-Resolution (ICCV 2019 · similarity 0.99)
- Forward Super-Resolution: How Can GANs Learn Hierarchical Generative Models for Real-World Distributions (ICLR 2023 · similarity 0.89)
- Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network (CVPR 2017 · similarity 0.88)
- Uncertainty-Aware GAN for Single Image Super Resolution (AAAI 2024 · similarity 0.86)
- SeD: Semantic-Aware Discriminator for Image Super-Resolution (CVPR 2024 · similarity 0.86)
- GSDD: Generative Space Dataset Distillation for Image Super-resolution (AAAI 2024 · similarity 0.86)
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.
