Problem
Research questions and friction points this paper is trying to address.
road segmentation
remote sensing images
interactive segmentation
fine-grained refinement
false positives/negatives
PC-SAM: Patch-Constrained Fine-Grained Interactive Road Segmentation in High-Resolution Remote Sensing Images
🤖 AI Summary
Existing fully automatic remote sensing road segmentation methods struggle with complex scenes and lack fine-grained, local interactive refinement capabilities, while general-purpose interactive models such as SAM perform poorly on remote sensing imagery. To address these limitations, this work proposes a unified framework that integrates fully automatic and interactive segmentation. Building upon the SAM architecture, the approach introduces a remote sensing–specific fine-tuning strategy and a Patch-Constrained mechanism that restricts the influence of point prompts to localized image patches. This enables precise interactive refinement of road masks in high-resolution remote sensing images. The method significantly outperforms existing fully automatic models across multiple remote sensing road datasets and, for the first time, supports efficient and flexible local fine-tuning, thereby enhancing segmentation accuracy and practicality in complex scenarios.
📝 Abstract
Road masks obtained from remote sensing images effectively support a wide range of downstream tasks. In recent years, most studies have focused on improving the performance of fully automatic segmentation models for this task, achieving significant gains. However, current fully automatic methods are still insufficient for identifying certain challenging road segments and often produce false positive and false negative regions. Moreover, fully automatic segmentation does not support local segmentation of regions of interest or refinement of existing masks. Although the SAM model is widely used as an interactive segmentation model and performs well on natural images, it shows poor performance in remote sensing road segmentation and cannot support fine-grained local refinement. To address these limitations, we propose PC-SAM, which integrates fully automatic road segmentation and interactive segmentation within a unified framework. By carefully designing a fine-tuning strategy, the influence of point prompts is constrained to their corresponding patches, overcoming the inability of the original SAM to perform fine local corrections and enabling fine-grained interactive mask refinement. Extensive experiments on several representative remote sensing road segmentation datasets demonstrate that, when combined with point prompts, PC-SAM significantly outperforms state-of-the-art fully automatic models in road mask segmentation, while also providing flexible local mask refinement and local road segmentation. The code will be available at https://github.com/Cyber-CCOrange/PC-SAM.
Innovation
Methods, ideas, or system contributions that make the work stand out.
interactive segmentation
patch-constrained prompting
fine-grained refinement
remote sensing road segmentation
SAM adaptation
🔎 Similar Papers
No similar papers found.
