A lightweight rail surface defect detection algorithm integrating multi-attention mechanism and content-aware reassembly of features

View Full Paper

Ask AI

Bookmark

Authors

Discussion

Enhanced lightweight detection model improves detection accuracy in rail inspections, indicating better safety outcomes.

MC-YOLOv8s improves mAP@0.5 by 14.4% and mAP@0.5:0.95 by 11.4%, enhancing rail surface defect detection capability.
Utilizing just 69.91% and 81.66% of the baseline model’s computational resources, it emphasizes efficiency and speed.
Development involves integrating multi-attention mechanisms with CARAFE for better micro-defect feature capture.
Results show a significant inference speed of 158 FPS, highlighting suitability for real-time railway operations.

Ma et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark