Research map / Image generation / Image restoration
HyperSIGMA: Hyperspectral Intelligence Comprehension Foundation Model
Di Wang, Meiqi Hu, Yao Jin, Yuchun Miao, Jiaqi Yang, Yichu Xu, Xiaolei Qin, Jiaqi Ma, Lingyu Sun, Chenxing Li, Chuan Fu, Hongruixuan Chen and 10 more
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 spectral subnetwork with spectral enhancement module to SpatSIGMA (ablation) | SpatSIGMA (ViT-Base) · Indian Pines classification | OA (%): 85.08 -> 85.54 | Improved |
| add spectral subnetwork with spectral enhancement module to SpatSIGMA (ablation) | SpatSIGMA (ViT-Base) · Houston classification | OA (%): 87.33 -> 86.80 | Hurt |
| add spectral subnetwork to SpatSIGMA (ablation) | SpatSIGMA (ViT-Base) · Xiongan classification | OA (%): 93.68 -> 95.06 | Improved |
| scale spatial ViT from Base to Huge (ablation) | SpatSIGMA · Indian Pines classification | OA (%): 85.08 -> 86.74 | Improved |
| scale spectral ViT from Base to Huge (spatial stays Base) (ablation) | HyperSIGMA · Indian Pines classification | OA (%): 85.54 -> 87.38 | Improved |
| scale both spatial and spectral ViT from Base to Huge (ablation) | HyperSIGMA · Xiongan classification | OA (%): 95.06 -> 95.78 | Improved |
| scale both ViTs from Base to Huge (ablation) | HyperSIGMA · Indian Pines classification | OA (%): 85.54 -> 87.70 | Improved |
| add spectral subnetwork under PGD attack (ablation) | SpatSIGMA · Indian Pines classification (PGD attack) | OA (%): 95.23 -> 95.92 | Improved |
| add spectral subnetwork (ablation) | SpatSIGMA · OSCD multispectral change detection | F1 (%): 58.53 -> 59.28 | Improved |
| add spectral subnetwork (ablation) | SpatSIGMA · Hermiston change detection | F1 (%): 92.08 -> 91.74 | Hurt |
| add spectral subnetwork (ablation) | SpatSIGMA · WDC Mall denoising (mixed noise) | PSNR: 28.103 -> 28.503 | Improved |
| add spectral subnetwork (ablation) | SpatSIGMA · Urban unmixing | abundance MSE: 0.0176 -> 0.0157 | Improved |
Later papers that compared with this method
1 later accepted paper reports HyperSIGMA as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- OneHSI: A Unified Hyperspectral Foundation Model with Physical Consistency (ECCV 2026 · better on 10 of 10 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
- ESSAformer: Efficient Transformer for Hyperspectral Image Super-resolution (ICCV 2023 · similarity 0.89)
- HSIGene: A Foundation Model for Hyperspectral Image Generation (TPAMI 2025 · similarity 0.88)
- LESSViT: Robust Hyperspectral Representation Learning under Spectral Configuration Shift (NeurIPS 2026 · similarity 0.88)
- MP-HSIR: A Multi-Prompt Framework for Universal Hyperspectral Image Restoration (ICCV 2025 · similarity 0.88)
- HSOD-BIT-V2: A Challenging Benchmark for Hyperspectral Salient Object Detection (AAAI 2025 · similarity 0.87)
- Automatic Spectral Calibration of Hyperspectral Images: Method, Dataset and Benchmark (CVPR 2025 · 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.
