Label-Efficient School Detection from Aerial Imagery via Weakly Supervised Pretraining and Fine-Tuning

2026-05-05

AI Summary

This study addresses the global scarcity of precise geospatial data on schools and the prohibitive cost of manual annotation in remote sensing imagery. To overcome these challenges, the authors propose a weakly supervised two-stage training framework: first, a model is pretrained using automatically generated school bounding boxes derived from sparse location points and semantic segmentation; then, it is fine-tuned with only 50 manually annotated images. By innovatively integrating automated pseudo-labeling with weak supervision, this approach establishes an end-to-end detection system that drastically reduces reliance on labeled data while achieving high-performance school identification. The method offers a scalable remote sensing solution for mapping educational infrastructure and connectivity worldwide, significantly lowering annotation costs without compromising accuracy.

Abstract

Accurate school detection is essential for supporting education initiatives, including infrastructure planning and expanding internet connectivity to underserved areas. However, many regions around the world face challenges due to outdated, incomplete, or unavailable official records. Manual mapping efforts, while valuable, are labor-intensive and lack scalability across large geographic areas. To address this, we propose a weakly supervised framework for school detection from aerial imagery that minimizes the need for human annotations while supporting global mapping efforts. Our method is specifically designed for low-data regimes, where manual annotations are extremely scarce. We introduce an automatic labeling pipeline that leverages sparse location points and semantic segmentation to generate infrastructure masks from which we generate bounding boxes. Using these automatically labeled images, we train our detectors on a first training stage to learn a representation of what schools look like, then using a small set of manually labeled images, we fine-tune the previously trained models on this clean dataset. This two stage training pipeline enables large-scale and strong detection in low-data setting of school infrastructure with minimal supervision. Our results demonstrate strong object detection performance, particularly in the low-data regime, where the models achieve promising results using only 50 manually labeled images, significantly reducing the need for costly annotations. This framework supports education and connectivity initiatives worldwide by providing an efficient and extensible approach to mapping schools from space. All models, training code and auto-labeled data will be publicly released to foster future research and real-world impact.

Problem

• school detection
• aerial imagery
• label efficiency
• weak supervision
• low-data regime

Innovation

• weakly supervised learning
• automatic labeling
• two-stage training
• low-data regime
• aerial imagery