Search papers, labs, and topics across Lattice.
This paper introduces ISRS-DETR, a novel interactive segmentation framework tailored for remote sensing imagery that leverages object-level evidence to enhance segmentation efficiency. By employing a detection-guided approach and a Dynamic Top-K Click Selection strategy, the framework allows a single user interaction to inform multiple same-class objects, significantly reducing the number of clicks required for satisfactory segmentation. Experimental results demonstrate that ISRS-DETR not only achieves state-of-the-art accuracy but also lowers the average number of clicks per image, addressing a critical barrier in practical applications of interactive segmentation in remote sensing.
A single click can now inform the segmentation of multiple objects in remote sensing images, drastically cutting down the annotation effort required.
Interactive segmentation reduces the prohibitive cost of pixel-level annotation by allowing users to delineate objects with a few clicks. However, applying this paradigm directly to remote sensing imagery is non-trivial: ultra-high resolutions, small object sizes, and sparse spatial distributions all degrade segmentation quality. Recent work has addressed the resolution barrier and achieved competitive results in interactive segmentation for remote sensing (ISRS). However, they treat all instances of a class within an image as a single objective target. Consequently, interactions spent on one object contribute nothing to its same-class neighbours, and satisfactory masks may demand up to 40 clicks per image, hindering the practicality of these frameworks. We observe that remote sensing scenes exhibit markedly strong inter-object correlation, meaning a single clicked object is highly informative about the rest of its category. Building on this, we propose ISRS-DETR, a detection-guided interactive segmentation framework that injects object-level evidence into both training and inference. Our ISRS-DETR employs an RF-DETR decoder with the interactive segmentation backbone to localise co-occurring same-class objects, and introduces a Dynamic Top-K Click Selection strategy that retains only reliable proposals and converts each into a simulated click, so one user interaction propagates across an entire class. Experiments on three standard remote sensing benchmarks show that ISRS-DETR achieves state-of-the-art accuracy while substantially reducing Number of Clicks per Image (NoC-I). All codes and data splits will be released for reproducibility upon acceptance.
