In this paper, we propose ReSeg-CLIP, a new training-free Open-Vocabulary Semantic Segmentation method for remote sensing data. To compensate for the problems of vision language models, such as CLIP in semantic segmentation caused by inappropriate interactions within the self-attention layers, we introduce a hierarchical scheme utilizing masks generated by SAM to constrain the interactions at multiple scales. We also present a model composition approach that averages the parameters of multiple RS-specific CLIP variants, taking advantage of a new weighting scheme that evaluates representational quality using varying text prompts. Our method achieves state-of-the-art results across three RS benchmarks without additional training.
@article{arxiv.2602.23869,
title = {Open-Vocabulary Semantic Segmentation in Remote Sensing via Hierarchical Attention Masking and Model Composition},
author = {Mohammadreza Heidarianbaei and Mareike Dorozynski and Hubert Kanyamahanga and Max Mehltretter and Franz Rottensteiner},
journal= {arXiv preprint arXiv:2602.23869},
year = {2026}
}
Comments
Published in the proceedings of the British Machine Vision Conference Workshops 2025