Research map / Image generation / Super-resolution
Image Super-Resolution Using Very Deep Residual Channel Attention Networks
Yulun Zhang, Kunpeng Li, Kai Li, Lichen Wang, Bineng Zhong, Yun Fu
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 long skip connection (LSC) only to the 200-block network without CA (ablation) | RCAN (no CA) · Set5 x2 | PSNR (dB): 37.45 -> 37.77 | Improved |
| add LSC and SSC (RIR) to the network without channel attention (ablation) | RCAN (no CA) · Set5 x2 | PSNR (dB): 37.45 -> 37.87 | Improved |
| add LSC and SSC (RIR) to the network with channel attention (ablation) | RCAN (with CA) · Set5 x2 | PSNR (dB): 37.52 -> 37.90 | Improved |
| add short skip connection (SSC) only without CA (ablation) | RCAN (no CA) · Set5 x2 | PSNR (dB): 37.45 -> 37.81 | Improved |
| add channel attention to the network without RIR (ablation) | RCAN (no RIR) · Set5 x2 | PSNR (dB): 37.45 -> 37.52 | Improved |
| add channel attention to the network with RIR (ablation) | RCAN (with RIR) · Set5 x2 | PSNR (dB): 37.87 -> 37.90 | Improved |
| use RCAN instead of EDSR (16M vs 43M parameters) at scale 8 | RCAN · Urban100 x8 | PSNR (dB): 22.51 -> 23.00 | Improved |
| use RCAN instead of EDSR at scale 8 | RCAN · Manga109 x8 | PSNR (dB): 24.69 -> 25.24 | Improved |
| add self-ensemble to RCAN | RCAN · Urban100 x4 | PSNR (dB): 26.82 -> 27.10 | Improved |
| use RCAN-upscaled images instead of FSRCNN for ResNet-50 recognition | RCAN · ImageNet subset recognition | top-1 error: 0.437 -> 0.393 | Improved |
| use RCAN instead of RDN at scale 4 | RCAN · Urban100 x4 | PSNR (dB): 26.61 -> 26.82 | Improved |
Closest papers by meaning
- Second-Order Attention Network for Single Image Super-Resolution (CVPR 2019 · similarity 0.92)
- Image Super-Resolution Using Deep Convolutional Networks (TPAMI 2016 · similarity 0.91)
- Accurate Image Super-Resolution Using Very Deep Convolutional Networks (CVPR 2016 · similarity 0.91)
- Single Image Super-Resolution via a Holistic Attention Network (ECCV 2020 · similarity 0.91)
- Fast, Accurate, and Lightweight Super-Resolution with Cascading Residual Network (ECCV 2018 · similarity 0.90)
- SAVSR: Arbitrary-Scale Video Super-Resolution via a Learned Scale-Adaptive Network (AAAI 2024 · 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.
