Research map / Detection and segmentation / CNN features
Beyond Euclidean Assumptions: Geometry-Aware Adaptive Routing for Remote Sensing Segmentation
Jie Qiu, Dizuo Cao, Linwei Dai, Xin Li, Fan Yang, Dong Yu, Changying Wang, Zongheng Wen, Youqin Chen, Jianzhang Chen
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 GAAR | UNetFormer ResNet18/34/50 · LoveDA | mIoU: +1.82, +2.22, +2.53 | Improved |
| add GAAR | LSKNet-T/S · LoveDA | mIoU: +1.93, +2.08 | Improved |
| add GAAR | DC-Swin Swin-T/S/B · LoveDA | mIoU: +1.24, +1.20, +1.21 | Improved |
| add FCW routing module (ablation) | GAAR on UNetFormer ResNet18 · LoveDA (UNetFormer ResNet18) | mIoU: 43.73 -> 52.80 | Improved |
| add Euclidean attention branch (ablation) | GAAR on UNetFormer ResNet18 · LoveDA (UNetFormer ResNet18) | mIoU: 50.71 -> 52.80 | Improved |
| add Hyperbolic attention branch (ablation) | GAAR on UNetFormer ResNet18 · LoveDA (UNetFormer ResNet18) | mIoU: 51.32 -> 52.80 | Improved |
| add GADR regularization (ablation) | GAAR on UNetFormer ResNet18 · LoveDA (UNetFormer ResNet18) | mIoU: 50.66 -> 52.80 | Improved |
| add adaptive Gumbel router (ablation) | GAAR on UNetFormer ResNet50 · LoveDA (UNetFormer ResNet50) | mIoU: 53.41 -> 54.05 | Improved |
| add temperature annealing (ablation) | GAAR on UNetFormer ResNet18 · LoveDA (UNetFormer ResNet18) | mIoU: 51.33 -> 52.80 | Improved |
| route hybrid rather than fully hyperbolic (ablation) | GAAR on U-Net · LoveDA | FPS: 7.45 -> 30.31 | Improved |
| hybrid routing vs GAAR HYP only (ablation) | GAAR on U-Net · LoveDA | mIoU: 51.22 -> 53.17 | Improved |
| raise GADR weight lambda from 0.005 to 0.1 (ablation) | GAAR on UNetFormer ResNet50 · LoveDA | mIoU: 54.05 -> 51.19 | Hurt |
Closest papers by meaning
- HySeg: Learning Generative Priors for Structure-Aware Remote Sensing Segmentation (CVPR 2026 · similarity 0.88)
- Sparse and Complete Latent Organization for Geospatial Semantic Segmentation (CVPR 2022 · similarity 0.87)
- UniGeoSeg: Towards Unified Open-World Segmentation for Geospatial Scenes (CVPR 2026 · similarity 0.87)
- SatSynth: Augmenting Image-Mask Pairs through Diffusion Models for Aerial Semantic Segmentation (CVPR 2024 · similarity 0.87)
- SemStereo: Semantic-Constrained Stereo Matching Network for Remote Sensing (AAAI 2025 · similarity 0.86)
- Towards Open-Vocabulary Remote Sensing Image Semantic Segmentation (AAAI 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.
