Research map / Image generation / Super-resolution
Efficient Residual Dense Block Search for Image Super-Resolution
Dehua Song, Chang Xu, Xu Jia, Yiyi Chen, Chunjing Xu, Yunhe Wang
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 |
|---|---|---|---|
| replace RDB with CRDB (pooling+recursion) (ablation) | light RDN · Urban100 x2 | PSNR: 31.94 -> 32.20 | Improved |
| add recursive operator to pooled CRDB (ablation) | CRDN · Urban100 x2 | PSNR: 31.90 -> 32.20 | Improved |
| pooling CRDB without recursion (ablation) | light RDN · Urban100 x2 | PSNR: 31.94 -> 31.90 | Hurt |
| pooling CRDB without recursion (ablation) | light RDN · x2 720p | Multi-Adds: 235.6 -> 163.9 | Improved |
| place CRDBs first or cross instead of last (ablation) | CRDN-O · Urban100 x2 | PSNR: 31.81 -> 31.90 | Improved |
| too many CRDBs (more than 5) (ablation) | SRDN · x2 | PSNR | Hurt |
| NAS over block types vs single block type (ablation) | efficient RDN · Urban100 x2 | PSNR: 32.20 -> 32.37 | Improved |
| credit-guided mutation instead of random mutation (ablation) | evolutionary search · Urban100 x2 | PSNR: 32.25 -> 32.37 | Improved |
| searched ESRN vs FALSR-A at similar cost | lightweight SR · Urban100 x2 | PSNR: 31.93 -> 32.37 | Improved |
Closest papers by meaning
- LatticeNet: Towards Lightweight Image Super-resolution with Lattice Block (ECCV 2020 · similarity 0.89)
- Fast, Accurate, and Lightweight Super-Resolution with Cascading Residual Network (ECCV 2018 · similarity 0.88)
- Multi-scale Residual Network for Image Super-Resolution (ECCV 2018 · similarity 0.88)
- Image Super-Resolution Using Dense Skip Connections (ICCV 2017 · similarity 0.88)
- Densely Residual Laplacian Super-Resolution (TPAMI 2020 · similarity 0.88)
- Fast and Accurate Single Image Super-Resolution via Information Distillation Network (CVPR 2018 · similarity 0.88)
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.
