A YOLO-Based Semi-Automated Labeling Approach to Improve Fault Detection Efficiency in Railroad Videos

2025-04-01

AI Summary

To address the high time consumption, error rate, and cost associated with manual annotation in railway video fault detection, this paper proposes a lightweight iterative YOLO-based semi-automatic annotation method. Starting from a small set of initial annotations, the method employs a closed-loop optimization cycle comprising model training, inference, and human correction. It supports both textual export of detection results and interactive, rapid annotation refinement. Unlike proprietary black-box commercial platforms, our open-source, lightweight solution ensures full transparency and ease of deployment. Experimental results show that per-image annotation time is reduced from 2–4 minutes to 30 seconds–2 minutes, while annotation error rates decrease significantly. Consequently, both human effort and temporal overhead are substantially reduced. The proposed framework establishes an efficient, scalable, and reproducible annotation paradigm for constructing large-scale railway fault detection datasets.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Semi-Supervised Learning
• Data Mining & Knowledge Management: Anomaly/Outlier Detection

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Semantics and Knowledge: Representation, semantic annotation, enhancement, enrichments, access and/or integration of a variety of data on the Web

Abstract

Manual labeling for large-scale image and video datasets is often time-intensive, error-prone, and costly, posing a significant barrier to efficient machine learning workflows in fault detection from railroad videos. This study introduces a semi-automated labeling method that utilizes a pre-trained You Only Look Once (YOLO) model to streamline the labeling process and enhance fault detection accuracy in railroad videos. By initiating the process with a small set of manually labeled data, our approach iteratively trains the YOLO model, using each cycle's output to improve model accuracy and progressively reduce the need for human intervention. To facilitate easy correction of model predictions, we developed a system to export YOLO's detection data as an editable text file, enabling rapid adjustments when detections require refinement. This approach decreases labeling time from an average of 2 to 4 minutes per image to 30 seconds to 2 minutes, effectively minimizing labor costs and labeling errors. Unlike costly AI based labeling solutions on paid platforms, our method provides a cost-effective alternative for researchers and practitioners handling large datasets in fault detection and other detection based machine learning applications.

Problem

• Reducing manual labeling time for railroad video datasets
• Improving fault detection accuracy with semi-automated YOLO
• Providing cost-effective labeling alternative for large datasets

Innovation

• YOLO-based semi-automated labeling for railroad videos
• Iterative training with minimal manual intervention
• Editable text export for rapid prediction correction