Research map / Image generation / Super-resolution
Invertible Image Rescaling
Mingqing Xiao, Shuxin Zheng, Chang Liu, Yaolong Wang, Di He, Guolin Ke, Jiang Bian, Zhouchen Lin, Tie-Yan Liu
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 |
|---|---|---|---|
| use L2 for LR guidance loss instead of L1 (ablation) | IRN 4x · Set5 4x | PSNR (dB): 34.75 -> 36.19 | Improved |
| use L2 guidance loss instead of L1 (with L2 reconstruction) (ablation) | IRN 4x · Urban100 4x | PSNR (dB): 30.79 -> 31.40 | Improved |
| use L1 instead of L2 for HR reconstruction loss (L2 guide) (ablation) | IRN 4x · Set5 4x | PSNR (dB): 35.93 -> 36.19 | Improved |
| use L1 instead of L2 for HR reconstruction loss (L2 guide) (ablation) | IRN 4x · Urban100 4x | PSNR (dB): 31.40 -> 31.41 | No clear effect |
| add partial distribution matching loss (ablation) | IRN 4x · Set14 4x | PSNR (dB): 32.18 -> 32.67 | Improved |
| add partial distribution matching loss (ablation) | IRN 4x · Urban100 4x | PSNR (dB): 30.91 -> 31.41 | Improved |
| add partial distribution matching loss (ablation) | IRN 4x · DIV2K val 4x | PSNR (dB): 34.90 -> 35.07 | Improved |
| add partial distribution matching loss (ablation) | IRN 4x · Set5 4x | SSIM: 0.9455 -> 0.9451 | No clear effect |
| replace Bicubic and EDSR super-resolution with IRN invertible downscaling and upscaling | Bicubic & SR · Set5 4x | PSNR (dB): 32.62 -> 36.19 | Improved |
| replace Bicubic and SAN with IRN at 2x | Bicubic & SR · Set5 2x | PSNR (dB): 38.31 -> 43.99 | Improved |
| use IRN+ (distribution-matching, perceptual) instead of IRN | IRN 4x · Set5 4x | PSNR (dB): 36.19 -> 33.59 | Hurt |
| sample a different z in IRN | IRN · benchmark average | PSNR (dB) | No clear effect |
Closest papers by meaning
- Invertible Rescaling Network and Its Extensions (IJCV 2022 · similarity 0.95)
- Self-Asymmetric Invertible Network for Compression-Aware Image Rescaling (AAAI 2023 · similarity 0.90)
- Plug-and-Play Tri-Branch Invertible Block for Image Rescaling (AAAI 2025 · similarity 0.89)
- Block-Based Multi-Scale Image Rescaling (AAAI 2025 · similarity 0.88)
- Deep Generative Model based Rate-Distortion for Image Downscaling Assessment (CVPR 2024 · similarity 0.86)
- Towards Realistic Data Generation for Real-World Super-Resolution (ICLR 2025 · similarity 0.84)
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.
