A Cross-Scale Sparse Attention Model (CSSAM), which aligns multiscale visual features with complex referring expressions through two key modules, and introduces M$^{3}RRSIS-D, a RRSIS dataset explicitly designed around three core reasoning dimensions: multi-target, reference-dependent multi-hop, and multi-granularity expressions.
Abstract
High-resolution remote sensing (RS) imagery provides fine spatial detail over large areas, making it essential for fine-grained scene understanding. Referring Remote Sensing Image Segmentation (RRSIS) aims to segment regions described by natural language, yet many existing methods rely on sequential alignment--fusion pipelines with limited joint modeling of cross-scale visual hierarchies and linguistic dependencies. Moreover, dense attention mechanisms struggle with the sparsity and wide spatial dispersion of RS targets, while current datasets offer limited structured evaluation of multi-target references, multi-hop relational reasoning, and multi-granularity descriptions, limiting both model development and evaluation. To address these challenges, we propose a Cross-Scale Sparse Attention Model (CSSAM), which aligns multiscale visual features with complex referring expressions through two key modules. The Cross-scale Modality-aligned Fusion Module (CMFM) performs unified cross-scale cross-modal fusion, while the Dynamic Multimodal Sparse Hybrid Attention (DMSHA) selectively attends to informative visual--textual tokens, reducing dense visual interactions. Furthermore, we introduce M$^{3}$RRSIS-D, a RRSIS dataset explicitly designed around three core reasoning dimensions: multi-target (M$^{1}$), reference-dependent multi-hop (M$^{2}$), and multi-granularity (M$^{3}$) expressions. Extensive experiments on M$^{3}$RRSIS-D and two public RRSIS benchmarks demonstrate that CSSAM achieves competitive performance across accuracy, fine-grained reasoning, and high-IoU localization metrics, validating its effectiveness for complex semantic reasoning in realistic RS scenarios. Code, dataset, and model weights will be released upon acceptance.
Results confirm that BMCA effectively preserves fine-grained spatial details, promotes semantic consistency across modalities, and provides a robust and generalizable solution for feature-level SAR–MS fusion in remote-sensing semantic segmentation.
Hao-Tian Xu, Xi-Min Yuan, Xiu-Jie Wang et al.· Remote Sensing· 0 citations
Few-shot semantic segmentation (FSSS) in remote sensing faces challenges due to limited annotated samples, significant domain shifts, and complex interclass variations. In this article, a cross-domain cooperative fusion network (CDCFNet) is proposed for few-shot remote sensing imagery segmentation across heterogeneous...
Hong-Yu Zhao, Yan-Qing Yao, Mao-Guo Gong et al.· IEEE Transactions on Geoscie...· 0 citations
This work proposes M2-DINO, a parameter-efficient multimodal adaptation framework for remote sensing semantic segmentation, where topology-oriented terminology refers to architectural sensitivity to topology-relevant structures, such as thin-object continuity and object boundaries, rather than explicit optimization of...
Zi-Fei Luo, Yu-Bin He, Zhi-Yong Li et al.· IEEE Journal of Selected Top...· 0 citations
DLPANet is proposed, a novel dual-level prototype alignment network centered on Prototype-Guided Spatial Attention, enabling simultaneous modeling of scene context and fine-grained details and demonstrates that the decoupled dual cross-attention mechanism provides superior prototype-query alignment compared to prior gl...
Mustafa Alawadi, M. Fateh· Jordanian Journal of Compute...· 0 citations
Referring Remote Sensing Image Segmentation (RRSIS) addresses the task of identifying ground objects guided by natural language descriptions. While Transformer-based architectures have shown promise, significant limitations remain in processing high-resolution imagery due to quadratic computational complexity and the l...
Shi-Miao Fan, Yi-Jie Zhang, Zhi-Ling Zhang et al.· International Conference on...· 0 citations
This paper establishes the first cross-domain RRSIS benchmark, designated as the Vaihingen-Potsdam Referring (VPRef) dataset, and develops a tailored parameter-efficient domain adaptation baseline anchored on the Segment Anything Model (SAM3) via Low-Rank Adaptation (LoRA).
Quan-Wei Liu, Tao Huang, Jia-Qi Yang et al.· arXiv.org· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.