Table 3.

The results of different numbers of edge supervised on our dataset. Bold black text represents the best results.

Methods	$Dice(\%)\uparrow$	$Sens.(\%)\uparrow$	$Prec.(\%)\uparrow$	$MAE(\%)\downarrow$	$E_{\varphi }$(%)$\uparrow$	$S_{\alpha }$(%)$\uparrow$
ResUNet_C${}_1$F	$89.16\pm 0.49$	$88.03\pm 1.49$	$\mathbf{92}.\mathbf{08}\pm \mathbf{1}.\mathbf{05}$	$0.56\pm 0.03$	$95.27\pm 0.12$	$85.59\pm 0.80$
ResUNet_C${}_2$F	$\mathbf{89}.\mathbf{93}\pm \mathbf{0}.\mathbf{09}$	$90.29\pm 0.66$	$91.91\pm 0.97$	$\mathbf{0}.\mathbf{52}\pm \mathbf{0}.\mathbf{01}$	$\mathbf{95}.\mathbf{69}\pm \mathbf{0}.\mathbf{10}$	$\mathbf{86}.\mathbf{75}\pm \mathbf{0}.\mathbf{07}$
ResUNet_C${}_3$F	$89.44\pm 0.14$	$90.15\pm 0.88$	$91.90\pm 1.13$	$0.55\pm 0.01$	$95.30\pm 0.49$	$85.41\pm 0.09$
ResUNet_C${}_4$F	$89.40\pm 0.33$	$\mathbf{90}.\mathbf{66}\pm \mathbf{0}.\mathbf{45}$	$91.12\pm 0.92$	$0.58\pm 0.02$	$95.32\pm 0.24$	$85.35\pm 0.43$
ResUNet_C${}_5$F	$88.33\pm 0.89$	$90.28\pm 0.67$	$90.05\pm 0.45$	$0.58\pm 0.06$	$95.05\pm 1.30$	$85.29\pm 1.58$