Automatic Labelling for Low-Light Pedestrian Detection

2025-07-03

AI Summary

Low-light RGB pedestrian detection is hindered by the scarcity of large-scale, accurately annotated datasets. To address this, we propose an end-to-end infrared (IR)-guided automatic annotation framework: a fine-tuned IR detector generates initial bounding boxes, which are then transferred to RGB images via cross-modal geometric and semantic alignment, yielding high-quality pseudo-labels on the KAIST multimodal dataset. Our pipeline surpasses human-annotated baselines on 6 out of 9 detection metrics—particularly achieving consistent gains in mAP@50 and mAP@50–95 on unseen sequences—demonstrating both the effectiveness and generalizability of pseudo-labels for low-light detection. To our knowledge, this is the first work enabling fully automatic, trainable IR-to-RGB label transfer without manual intervention. The source code is publicly available.

Technology Category

• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Computer Vision: Multi-modal Vision
• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification

Application Category

• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Pedestrian detection in RGB images is a key task in pedestrian safety, as the most common sensor in autonomous vehicles and advanced driver assistance systems is the RGB camera. A challenge in RGB pedestrian detection, that does not appear to have large public datasets, is low-light conditions. As a solution, in this research, we propose an automated infrared-RGB labeling pipeline. The proposed pipeline consists of 1) Infrared detection, where a fine-tuned model for infrared pedestrian detection is used 2) Label transfer process from the infrared detections to their RGB counterparts 3) Training object detection models using the generated labels for low-light RGB pedestrian detection. The research was performed using the KAIST dataset. For the evaluation, object detection models were trained on the generated autolabels and ground truth labels. When compared on a previously unseen image sequence, the results showed that the models trained on generated labels outperformed the ones trained on ground-truth labels in 6 out of 9 cases for the mAP@50 and mAP@50-95 metrics. The source code for this research is available at https://github.com/BouzoulasDimitrios/IR-RGB-Automated-LowLight-Pedestrian-Labeling

Problem

• Automates labeling for low-light pedestrian detection
• Transfers labels from infrared to RGB images
• Improves detection accuracy in low-light conditions

Innovation

• Automated infrared-RGB labeling pipeline
• Infrared detection with fine-tuned model
• Label transfer from infrared to RGB