The Improvement of Automated Crack Segmentation on Concrete Pavement with Graph Network

Table 1

The comparison of the test results on the concrete pavement image dataset.

	Precision	Recall	F1	IoU
AutoEncoder	0.38	0.53	0.42	0.28
PSPNet [37]	0.65	0.44	0.52	0.35
U-Net [19]	0.67	0.37	0.45	0.31
KiU-Net [38]	0.38	0.69	0.46	0.31
GA-Unet(ours)	0.63	0.49	0.53	0.37