Active-SAOOD: Active Sparsely Annotated Oriented Object Detection in Remote Sensing Images

2026-05-11

AI Summary

This work addresses the high annotation cost of oriented object detection in remote sensing imagery and the limitations of existing sparse annotation methods, which rely on predefined class distributions and inadequately model the characteristics of sparse samples. To overcome these challenges, the paper proposes a novel active learning–based approach for sparse annotation in oriented object detection. It introduces, for the first time, active learning to this task by incorporating a model state observation module that dynamically selects high-value instances for annotation based on a comprehensive assessment of orientation, classification, and localization uncertainties, as well as intra- and inter-class diversity—all without requiring prior knowledge of class distributions. Experimental results demonstrate that with only 1% of labeled data, the proposed method outperforms baseline approaches by up to 9% across multiple remote sensing benchmarks, significantly enhancing training stability and practicality under sparse annotation settings.

Abstract

Reducing the annotation cost of oriented object detection in remote sensing remains a major challenge. Recently, sparse annotation has gained attention for effectively reducing annotation redundancy in densely remote sensing scenes. However, (1) the sparse data reliance on class-dependent sampling, and (2) the lack of in-depth investigation into the characteristics of sparse samples hinders its further development. This paper proposes an active learning-based sparsely annotated oriented object detection (SAOOD) method, termed Active-SAOOD. Based on a model state observation module, Active-SAOOD actively selects the most valuable sparse samples at the instance level that are best suited to the current model state, by jointly considering orientation, classification, and localization uncertainty, as well as inter- and intra-class diversity. This design enables SAOOD to operate stably under completely randomly initialized sparse annotations and extends its applicability to broader real-world. Experiments on multiple datasets demonstrate that Active-SAOOD significantly improves both performance and stability of existing SAOOD methods under various random sparse annotation. In particular, with only 1\% annotated ratios, it achieves a 9\% performance gain over the baseline, further enhancing the practical value of SAOOD in remote sensing. The code will be public.

Problem

• oriented object detection
• sparse annotation
• annotation cost
• remote sensing images
• active learning

Innovation

• active learning
• sparse annotation
• oriented object detection
• uncertainty estimation
• remote sensing