Selective Prototype Aggregation for Remote Sensing Few-Shot Semantic Segmentation

Kun Liu

doi:10.1109/TGRS.2025.3612893

Selective Prototype Aggregation for Remote Sensing Few-Shot Semantic Segmentation

2025-01-01

0

PRE

AI

DOI：10.1109/TGRS.2025.3612893

Original

Save

Abstract

En 中文

Few-shot semantic segmentation (FSS) in remote sensing imagery remains challenging due to complex scenes, large image sizes, frequent multiclass coexistence, and significant intraclass variation. These factors often lead to severe overfitting to base classes and hinder generalization to novel classes. To address these issues, we propose a Mamba-enhanced transformer (MET) framework with selective prototype aggregation (SPA), which simultaneously suppresses base class interference and enhances novel class representation. Specifically, the dynamic feature aggregation (DFA) module integrates features from a base class learner and a metalearner, effectively enhancing interclass distinction and mitigating feature contamination in multiclass coexistence scenarios. We further introduce the MET, which combines Mamba blocks and Transformer self-attention to capture long-range dependencies and multiscale context while maintaining low computational complexity. Mamba blocks independently extract support and query features, preventing cross-branch interference and improving robustness. By selectively aggregating support prototypes with query self-support prototypes, SPA strengthens feature discriminability and leverages query-specific contextual cues to reduce intraclass variance. Our model demonstrates strong performance across multiple datasets, achieves 50.49% mean intersection over union (mIoU) on the iSAID dataset and 30.82% on the LoveDA dataset, representing improvements of 3.95% and 3.89%, respectively, over the base model. Ablation experiments further validate the superior performance of SPA.

Keywords:

Few-shot learning

few-shot semantic segmentation

prototype learning

remote sensing

semantic segmentation

Selective Prototype Aggregation for Remote Sensing Few-Shot Semantic Segmentation

Abstract

Journal

Organization

Cited Papers