GeoSeg: Training-Free Reasoning-Driven Segmentation in Remote Sensing Imagery
GeoSeg addresses the problem of segmentation in remote sensing imagery for researchers and practitioners in the field of remote sensing, providing an incremental solution to existing challenges.
GeoSeg tackles the problem of segmentation in remote sensing imagery, achieving zero-shot, training-free segmentation with a framework that couples MLLM reasoning with precise localization, outperforming all baselines. The framework is tested on GeoSeg-Bench, a diagnostic benchmark of 810 image-query pairs.
Recent advances in MLLMs are reframing segmentation from fixed-category prediction to instruction-grounded localization. While reasoning based segmentation has progressed rapidly in natural scenes, remote sensing lacks a generalizable solution due to the prohibitive cost of reasoning-oriented data and domain-specific challenges like overhead viewpoints. We present GeoSeg, a zero-shot, training-free framework that bypasses the supervision bottleneck for reasoning-driven remote sensing segmentation. GeoSeg couples MLLM reasoning with precise localization via: (i) bias-aware coordinate refinement to correct systematic grounding shifts and (ii) a dual-route prompting mechanism to fuse semantic intent with fine-grained spatial cues. We also introduce GeoSeg-Bench, a diagnostic benchmark of 810 image--query pairs with hierarchical difficulty levels. Experiments show that GeoSeg consistently outperforms all baselines, with extensive ablations confirming the effectiveness and necessity of each component.