Semi-supervised Source Detection in Astronomical Images: New Benchmark and Strong Baseline

2026-06-08

AI Summary

This study addresses the challenges posed by dense stellar sources, low signal-to-noise ratios, and scarce annotations in astronomical images, which severely limit the performance of existing object detection methods. To this end, the authors introduce LAMOST-DET, the first large-scale semi-supervised benchmark for astronomical detection, comprising 18,400 images and 728,898 annotated source instances. They further propose the Nova Teacher framework, which integrates a dual-teacher architecture, source-aware augmentation, confidence-guided pseudo-labeling, and a cross-view complementary mining mechanism. Evaluated on LAMOST-DET, the method achieves significant improvements of 4.04% and 5.22% in mean average precision (mAP) under two semi-supervised settings, respectively, and demonstrates strong generalization capability on natural image datasets.

Abstract

Source detection in modern observational astronomy is a cornerstone for localizing and identifying stellar sources accurately. It is crucial for studies such as stellar population synthesis and cosmological parameter estimation. However, the characteristics of astronomical images, including high density, the effect of point spread functions and low signal-to-noise ratios, significantly challenge the latest advanced object detectors. Besides, fully-supervised detection methods are hardly practical, due to the significant difficulty in annotating dense, small, and faint sources in astronomical images. To tackle the scarcity of astronomical datasets, we introduce a new comprehensive benchmark (LAMOST-DET), comprising 18,400 astronomical images and 728,898 source instances. Upon the dataset, we further devise a novel semi-supervised learning framework coined Nova Teacher, capable of detecting dense sources effectively given sparse annotations. It integrates source light enhancement module, confidence-guided pseudo-supervision, and cross-view complementary mining in a dual-teacher paradigm. Extensive experiments on LAMOST-DET show that, Nova Teacher consistently improves previous competitors by 4.04% and 5.22% mAP under two semi-supervised settings. Additionally, our method competes against other detectors on a natural image dataset, validating its generalization ability to various scenarios. The source code is available at https://github.com/AcWiz/NovaTeacher.

Problem

• source detection
• astronomical images
• semi-supervised learning
• low signal-to-noise ratio
• dense object detection

Innovation

• semi-supervised learning
• source detection
• astronomical images
• pseudo-supervision
• dual-teacher paradigm