Research map / Image generation / Super-resolution
Cross-Scale Internal Graph Neural Network for Image Super-Resolution
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 cross-scale graph aggregation (IGNN) to EDSR | EDSR · Urban100 x2 | PSNR: 32.93 -> 33.23 | Helped |
| IGNN vs EDSR | EDSR · Urban100 x4 | PSNR: 26.64 -> 26.84 | Helped |
| IGNN vs SAN | SISR · Set5 x4 | PSNR: 32.64 -> 32.57 | Hurt |
| add self-ensemble (IGNN+) | IGNN · Urban100 x2 | PSNR: 33.23 -> 33.42 | Helped |
| add same-scale non-local block (ablation) | EDSR · Urban100 x2 | PSNR: 32.93 -> 32.98 | Helped |
| use cross-scale instead of same-scale kNN aggregation (ablation) | IGNN · Urban100 x2 | PSNR: 33.01 -> 33.23 | Helped |
| place GraphAgg after 16th block vs 8th (ablation) | IGNN · Urban100 x2 | PSNR: 33.17 -> 33.23 | Helped |
| enlarge search window d from 10 to 30 (ablation) | IGNN · Urban100 x2 | PSNR: 33.14 -> 33.23 | Helped |
| increase neighbors k from 1 to 5 (ablation) | IGNN · Urban100 x2 | PSNR: 33.17 -> 33.23 | Helped |
| increase k beyond 5 (ablation) | IGNN · Urban100 x2 | PSNR: 33.23 -> 33.22 | No clear effect |
| add AdaPN (ablation) | IGNN · Urban100 x2 | PSNR: 33.18 -> 33.23 | Helped |
| add edge-conditioned sub-network (ablation) | IGNN · Urban100 x2 | PSNR: 33.12 -> 33.23 | Helped |
Closest papers by meaning
- Image Super-Resolution With Cross-Scale Non-Local Attention and Exhaustive Self-Exemplars Mining (CVPR 2020 · similarity 0.89)
- Non-Local Recurrent Network for Image Restoration (NeurIPS 2018 · similarity 0.87)
- Exploring the Low-Pass Filtering Behavior in Image Super-Resolution (ICML 2024 · similarity 0.86)
- Low-Res Leads the Way: Improving Generalization for Super-Resolution by Self-Supervised Learning (CVPR 2024 · similarity 0.86)
- Image Processing GNN: Breaking Rigidity in Super-Resolution (CVPR 2024 · similarity 0.86)
- One-Step Effective Diffusion Network for Real-World Image Super-Resolution (NeurIPS 2024 · similarity 0.85)
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.