Asymmetric semantic spatial learning improves remote sensing domain generalization
ASCoR: Asymmetric Semantic-Spatial Learning for Remote Sensing Domain Generalization
Computer Vision and Pattern Recognition
Summary
Remote sensing images come from different places and conditions, so it's hard for computer programs to understand them all well. The authors propose ASCoR, a method that learns both the meaning of categories from their names and the spatial patterns in images, then combines these to make better predictions on new, unseen areas. ASCoR is tested on several benchmarks and consistently performs better than prior methods. This approach helps computers better recognize land features in satellite images even when they haven’t been trained on data from that exact place.
What this means in practice
- •For remote sensing analysts: Produce more reliable land cover maps for new geographic areas without needing new training data for each location.
- •For gis software developers: Integrate ASCoR’s framework to enhance semantic segmentation modules that generalize better across different environmental domains.
- •For environmental monitoring companies: Use improved domain generalization techniques to offer more accurate and scalable satellite image analysis services across diverse regions.$Commercial implications: Enables scalable remote sensing data analysis products with higher accuracy in multiple domains for environmental clients.
Authors
Zitong Wang, Shuchang Lyu, Meng Li, Qiang Wang, Guangliang Cheng, Zhenwei Shi
Abstract
Source-only domain generalization (DG) for remote sensing semantic segmentation seeks reliable predictions in unseen domains without target-domain data during training. Although vision foundation models offer transferable knowledge, category recognition remains sensitive to source-specific visual cues, while task-specific adaptation may weaken pretrained transferability. We propose ASCoR, an asymmetric semantic-spatial framework built on SAM3. Its text-conditioned semantic branch uses class-name prompts to provide explicit category guidance. Probe-Once Dynamic Adaptive LoRA distributes a fixed adaptation budget across encoder blocks to limit source-specific specialization. A complementary task-specific dense branch learns multi-scale spatial representations using a separate frozen SAM3 image encoder. A lightweight reliability router integrates the two branches through bounded spatial correction based on local prediction reliability. All components are trained using source-domain supervision only. Experiments on the ISPRS, LoveDA, and OpenEarthMap benchmarks cover nine cross-domain transfers. ASCoR achieves a state-of-the-art mean IoU of 68.61% across six ISPRS transfers, outperforming the previous best method by 2.69 percentage points. It also achieves the best results on all three LoveDA/OpenEarthMap transfers, with 59.07%, 47.95%, and 49.38% mIoU on Rural-to-Urban, Urban-to-Rural, and Urban-to-OpenEarthMap, respectively. Ablation studies, feature visualizations, and spatial reliability analyses support the effectiveness and complementarity of the proposed components. Code: https://github.com/WangZitong42/ASCoR.