Story perspectives

YOLO-DGVG Model Revolutionizes Road Crack Detection Accuracy

11/1/2025

Summary

Story summary

The YOLO-DGVG model improves road crack detection over the original YOLOv8s, with higher accuracy and fewer parameters.
It outperformed other object detection networks, achieving best performance in eight of nine categories on the PID road crack dataset.
The dataset included pavement defects, and YOLO-DGVG localized and classified cracks, especially fine ones.
Potential biases in the training datasets may affect generalization to different road conditions.

1 of 1

Story summary

The YOLO-DGVG model improves road crack detection over the original YOLOv8s, with higher accuracy and fewer parameters.
It outperformed other object detection networks, achieving best performance in eight of nine categories on the PID road crack dataset.
The dataset included pavement defects, and YOLO-DGVG localized and classified cracks, especially fine ones.
Potential biases in the training datasets may affect generalization to different road conditions.