Research map / Image generation / Image restoration
Spatial-Spectral Residuals Informed Diffusion Neural Operator for Pan-sharpening
Jiahan Huang, Ran Ran, Junming Hou, Zihao Chen, Xiaofeng Cong, Junling Li, Liang-Jian Deng
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 pixel-wise spatial and spectral consistency residuals as denoiser inputs (ablation) | SRINO · WV3 (reduced resolution) | PSNR: 39.013 -> 39.305 | Improved |
| add spatial consistency residual only (ablation) | SRINO · WV3 (reduced resolution) | PSNR: 39.013 -> 39.177 | Improved |
| add spectral consistency residual only (ablation) | SRINO · WV3 (reduced resolution) | PSNR: 39.013 -> 39.124 | Improved |
| add spatial residual on top of spectral residual (both) (ablation) | SRINO · WV3 (reduced resolution) | SAM: 2.945 -> 2.869 | Improved |
| use residual-informed inputs instead of gradient-guided strategy (ablation) | SRINO · WV3 (reduced resolution) | PSNR, SAM, ERGAS, Q8 | Improved |
| vary gradient guidance weight alpha (1, 10, 100) in gradient-guided baseline (ablation) | gradient-guided strategy · WV3 (reduced resolution) | PSNR, SAM, ERGAS, Q8 | No clear effect |
| use Galerkin-type neural operator backbone instead of CNN (ablation) | SRINO denoiser · WV3 (reduced resolution) | PSNR, SAM, Q8, ERGAS | Improved |
| use Galerkin-type neural operator backbone instead of FNO (ablation) | SRINO denoiser · WV3 (reduced resolution) | PSNR, SAM, Q8, ERGAS | Improved |
| replace attention-based diffusion denoiser with Galerkin neural operator | PanDiff-style attention denoiser · efficiency (FLOPs, memory, inference time) | FLOPs, memory, inference time | Improved |
| SRINO vs ADWM on spatial distortion | WV3 full resolution · WV3 (full resolution) | Ds: 0.029 (ADWM) vs 0.032 (SRINO) | Mixed |
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.
- Fourmer — Fourmer: An Efficient Global Modeling Paradigm for Image Restoration (ICML 2023) · better on 1 of 1 shared result
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
- NODiff: Neural Operator Diffusion for Multispectral Image Fusion (AAAI 2026 · similarity 0.93)
- SSDiff: Spatial-spectral Integrated Diffusion Model for Remote Sensing Pansharpening (NeurIPS 2024 · similarity 0.91)
- Hyperspectral Pansharpening via Diffusion Models with Iteratively Zero-Shot Guidance (CVPR 2025 · similarity 0.89)
- Accelerated Diffusion via High-Low Frequency Decomposition for Pan-Sharpening (AAAI 2025 · similarity 0.89)
- Binarized Neural Network for Multi-spectral Image Fusion (CVPR 2025 · similarity 0.87)
- Unsupervised Pan-Sharpening via Mutually Guided Detail Restoration (AAAI 2024 · similarity 0.86)
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.
