Research map / Image generation / Super-resolution
OSRT: Omnidirectional Image Super-Resolution With Distortion-Aware Transformer
Fanghua Yu, Xintao Wang, Mingdeng Cao, Gen Li, Ying Shan, Chao Dong
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 distortion-aware OSRT instead of SwinIR | ODI-SR and SUN 360 Panorama under Fisheye downsampling, ×2 and ×4 · ODI-SR; SUN 360 Panorama | PSNR: 30.64 -> 30.77; 31.33 -> 31.52; 27.31 -> 27.41; 27.71 -> 27.84 | Improved |
| use OSRT rather than SphereSR | ODI-SR and SUN 360 Panorama under ERP downsampling · ODI-SR; SUN 360 Panorama ×8 and ×16 | WS-PSNR: 24.37 -> 24.53; 24.17 -> 24.38; 22.51 -> 22.69; 21.95 -> 22.13 | Improved |
| add DACB (ablation) | OSRT-light · ODI-SR ×2; SUN360 ×2 | PSNR: 30.27 -> 30.41; 30.78 -> 31.00 | Improved |
| add DAAB without window condition Cw (ablation) | OSRT-light · ODI-SR ×2; SUN360 ×2 | PSNR: 30.27 -> 30.31; 30.78 -> 30.83 | Improved |
| add window condition Cw to DAAB (ablation) | OSRT-light · ODI-SR ×2; SUN360 ×2 | PSNR: 30.31 -> 30.32; 30.83 -> 30.84 | No clear effect |
| combine DACB and DAAB with Cw (ablation) | OSRT-light · ODI-SR ×2; SUN360 ×2 | PSNR: 30.27 -> 30.44; 30.78 -> 31.04 | Improved |
| add DF2K-ERP pseudo-ERP data jointly (ablation) | SwinIR · SUN360 ×2 and ×4 | PSNR: 31.21 -> 31.33; 27.39 -> 27.71 | Improved |
| add DF2K plain-image data (ablation) | SwinIR · SUN360 ×2 and ×4 | PSNR: 31.21 -> 31.26; 27.39 -> 27.59 | Improved |
| train jointly rather than pretrain then fine-tune on DF2K-ERP and ODI-SR (ablation) | SwinIR · SUN360 ×2 | PSNR: 31.17 -> 31.33 | Improved |
| train jointly rather than pretrain then fine-tune on DF2K-ERP and ODI-SR (ablation) | SwinIR · SUN360 ×4 | PSNR: 27.74 -> 27.71 | Hurt |
Later papers that compared with this method
5 later accepted papers report OSRT as a baseline in their results tables; 5 of them report a better number on most of the results they share with it.
- D²R²OSR: Degradation-Disentangled Representation for Real-World Omnidirectional Image Super-Resolution (ECCV 2026 · better on 4 of 4 shared results)
- Edge-Focused Super-Resolution for Omnidirectional Images with Spherical Geometric Augmentation (CVPR 2026 · better on 2 of 2 shared results)
- Fast Omni-Directional Image Super-Resolution: Adapting the Implicit Image Function with Pixel and Semantic-Wise Spherical Geometric Priors (AAAI 2025 · better on 5 of 7 shared results)
- Omnidirectional Image Super-resolution via Bi-projection Fusion (AAAI 2024 · better on 6 of 6 shared results)
- OmniSSR: Zero-shot Omnidirectional Image Super-Resolution using Stable Diffusion Model (ECCV 2024 · better on 2 of 2 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- SphereSR — SphereSR: 360deg Image Super-Resolution With Arbitrary Projection via Continuous Spherical Image Representation (CVPR 2022) · better on 4 of 4 shared results
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
- Omnidirectional Image Super-resolution via Bi-projection Fusion (AAAI 2024 · similarity 0.91)
- Spherical Pseudo-Cylindrical Representation for Omnidirectional Image Super-resolution (AAAI 2024 · similarity 0.88)
- Fast Omni-Directional Image Super-Resolution: Adapting the Implicit Image Function with Pixel and Semantic-Wise Spherical Geometric Priors (AAAI 2025 · similarity 0.86)
- Occlusion-Embedded Hybrid Transformer for Light Field Super-Resolution (AAAI 2025 · similarity 0.84)
- OSFFNet: Omni-Stage Feature Fusion Network for Lightweight Image Super-Resolution (AAAI 2024 · similarity 0.83)
- OmniSplat: Taming Feed-Forward 3D Gaussian Splatting for Omnidirectional Images with Editable Capabilities (CVPR 2025 · similarity 0.83)
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.
