EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Pixel’s Neighbors Are Noteworthy: Localized Vision–Language Attention for Remote Sensing Semantic Segmentation

Cheng Zeng, Sheng Tao, Xiaowei Tan, Zhifeng Xiao, Lei Hu

Remote Sensing · 26.05.2026 · 18 (11) · 1708

DOI: 10.3390/rs18111708 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Remote Sensing, online erschienen am 26.05.2026. Verfasst von Cheng Zeng und 4 weiteren Autor:innen. Die Arbeit nennt 27 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University, Hubei Provincial Key Research and Development Program, Hubei Provincial Major Science and Technology Program.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

In recent years, vision–language models (VLMs) have been introduced into remote sensing semantic segmentation to provide richer semantic representations through visual–textual alignment. However, most existing VLM-based segmentation methods focus on global semantic alignment while neglecting pixel-level local neighborhood features, which are crucial for reliably understanding remote sensing imagery with high spatial resolution, complex structures, and strong spatial continuity. To address this issue, we propose LoVLANet (Localized Vision–Language Attention Network), a novel vision–language segmentation framework that integrates language-driven global semantics with local spatial context. LoVLANet consists of a text encoder, a visual encoder, and a segmentation decoder. Specifically, the text encoder is inherited from RemoteCLIP to preserve domain-adapted vision–language alignment. The visual encoder is built upon a Vision Transformer (ViT). To enhance local dependency modeling, we propose a Neighborhood Key–Key Encoder. It leverages a Gaussian-weighted neighborhood matrix for spatial correlation and uses key–key similarity to emphasize intrinsic semantic similarity over query-driven features, thus, preserving spatial consistency. Finally, the segmentation decoder fuses multi-scale visual features and aligns the image–text representations to generate accurate pixel-level segmentation results. Experiments on RGB remote sensing benchmarks, including LoveDA and GID, show that LoVLANet achieves competitive segmentation performance under the adopted experimental settings, with improved mIoU and clearer boundary delineation in qualitative visualizations. These results suggest the effectiveness of explicitly modeling local neighborhood relationships in VLM-based segmentation for supervised remote sensing scene understanding.

Förderung

  • State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University
  • Hubei Provincial Key Research and Development Program
  • Hubei Provincial Major Science and Technology Program
  • Wuhan Natural Science Foundation Focused Program

Zitieren

APA

Zeng, C., Tao, S., Tan, X., Xiao, Z., & Hu, L. (2026). Pixel’s Neighbors Are Noteworthy: Localized Vision–Language Attention for Remote Sensing Semantic Segmentation. Remote Sensing, 18(11), 1708. https://doi.org/10.3390/rs18111708

Vancouver

Zeng C, Tao S, Tan X, Xiao Z, Hu L. Pixel’s Neighbors Are Noteworthy: Localized Vision–Language Attention for Remote Sensing Semantic Segmentation. Remote Sensing. 2026;18(11):1708. doi:10.3390/rs18111708

BibTeX
@article{Zeng2026,
  author = {Zeng, Cheng and Tao, Sheng and Tan, Xiaowei and Xiao, Zhifeng and Hu, Lei},
  title = {Pixel’s Neighbors Are Noteworthy: Localized Vision–Language Attention for Remote Sensing Semantic Segmentation},
  journal = {Remote Sensing},
  year = {2026},
  volume = {18},
  number = {11},
  pages = {1708},
  doi = {10.3390/rs18111708}
}
Kostenlos starten →