Railway Subgrade Defect Automatic Recognition Method Based on Improved Faster R-CNN

Table 5

The results of the model performance comparison experiment.

Method	Detect speed/ (sec)		Precision	Recall	F-Score
	GPU	CPU
HOG+SVM	-	0.136	54.7%	29.8%	38.6%
Faster-RCNN	0.091	0.367	68.3%	66.9%	67.6%
The proposed method	0.093	0.391	85.2%	82.1%	83.6%