Research map / Detection and segmentation / CNN features
HySeg: Learning Generative Priors for Structure-Aware Remote Sensing Segmentation
Jie Qiu, Xin Li, Fan Yang, Yan Wang, Dong Yu, Changying Wang, Linwei Dai, Yongxiang Chen, 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 HySeg to backbone | UNetFormer ResNet18 · LoveDA | mIoU: 52.4 -> 54.5 | Improved |
| add HySeg to backbone | D2LS ConvNeXt B · LoveDA | mIoU: 55.3 -> 55.6 | Improved |
| add HySeg to backbone | LSKNet S · UAVid | mIoU: 70.0 -> 70.9 | Improved |
| add HySeg to backbone | UNetFormer ResNet18 · Potsdam topology | beta0 error: 1.6448 -> 1.2369 | Improved |
| add DAS head without prior (ablation) | UNetFormer ResNet18 · LoveDA | mIoU: 52.42 -> 52.76 | No clear effect |
| add simple prior fusion (ablation) | UNetFormer ResNet18 · LoveDA | mIoU: 52.42 -> 53.42 | Improved |
| add MeanStruct prior with DAS but no P2A (ablation) | UNetFormer ResNet18 · LoveDA | mIoU: 52.42 -> 52.01 | Hurt |
| add P2A and affinity message passing (ablation) | UNetFormer ResNet18 · LoveDA | mIoU: 53.53 -> 54.02 (w/o self-attn) / 54.48 full | Improved |
| add self-attention to DAS (ablation) | UNetFormer ResNet18 · LoveDA | mIoU: 54.02 -> 54.48 | Improved |
| replace discriminative priors (ResNet18/Swin B/LSKNet S/ConvNeXt B) with MeanStruct prior (ablation) | HySeg ResNet18 · LoveDA | mIoU: 53.87 (best other prior) -> 54.48 | Improved |
| use ConvNeXt B prior with ResNet18 backbone (ablation) | HySeg · LoveDA | mIoU: 52.42 -> 53.16 | No clear effect |
| learnable sigma instead of fixed sigma (ablation) | P2A · LoveDA | mIoU: 53.82 (best fixed 0.5) -> 54.48 | Improved |
Closest papers by meaning
- A Gift From the Integration of Discriminative and Diffusion-Based Generative Learning: Boundary Refinement Remote Sensing Semantic Segmentation (TPAMI 2026 · similarity 0.91)
- SatSynth: Augmenting Image-Mask Pairs through Diffusion Models for Aerial Semantic Segmentation (CVPR 2024 · similarity 0.89)
- CrossEarth: Geospatial Vision Foundation Model for Domain Generalizable Remote Sensing Semantic Segmentation (TPAMI 2025 · similarity 0.88)
- Sparse and Complete Latent Organization for Geospatial Semantic Segmentation (CVPR 2022 · similarity 0.88)
- Towards Open-Vocabulary Remote Sensing Image Semantic Segmentation (AAAI 2025 · similarity 0.87)
- ROS-SAM: High-Quality Interactive Segmentation for Remote Sensing Moving Object (CVPR 2025 · similarity 0.87)
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.
