Pavement Crack Detection and Segmentation Method Based on Improved Deep Learning Fusion Model

Table 2

Comparison of crack classification detection network prediction results.

Crack type	Precision		Recall		Average precision (AP)
	Before	After	Before	After	Before	After
Transverse	0.773	0.805	0.780	0.811	0.776	0.810
Longitudinal	0.771	0.827	0.782	0.836	0.779	0.834
Alligator	0.694	0.811	0.721	0.824	0.718	0.825
Mean average precision (mAP)					0.758	0.823