Table 2.

Comparisons between different networks on our dataset. Bold black text represents the best results.

Methods	$Dice(\%)\uparrow$	$Sens.(\%)\uparrow$	$Prec.(\%)\uparrow$	$MAE(\%)\downarrow$	$E_{\varphi }$(%)$\uparrow$	$S_{\alpha }$(%)$\uparrow$
U-Net [11]	$85.56\pm 0.33$	$85.38\pm 1.53$	$85.76\pm 0.89$	$0.72\pm 0.01$	$94.21\pm 0.02$	$81.23\pm 0.23$
UNet+ [9]	$86.71\pm 1.25$	$90.27\pm 0.61$	$88.30\pm 1.05$	$0.60\pm 0.02$	$94.50\pm 0.63$	$84.61\pm 1.00$
Attention U-Net [34]	$87.40\pm 0.26$	$89.48\pm 0.49$	$89.88\pm 0.53$	$0.58\pm 0.03$	$94.74\pm 0.87$	$84.71\pm 0.87$
PSPNet [17]	$87.45\pm 0.31$	$88.32\pm 1.25$	$89.89\pm 1.11$	$0.60\pm 0.05$	$93.84\pm 0.27$	$83.81\pm 0.20$
Deeplabv3 [18]	$87.81\pm 0.19$	$89.24\pm 0.96$	$90.72\pm 0.66$	$0.58\pm 0.02$	$95.58\pm 0.22$	$86.03\pm 0.95$
Inf-Net [6]	$88.49\pm 0.17$	$90.07\pm 0.35$	$90.39\pm 0.18$	$0.55\pm 0.01$	$\mathbf{95}.\mathbf{70}\pm \mathbf{0}.\mathbf{24}$	$86.55\pm 0.09$
SCRN [22]	$86.24\pm 0.08$	$83.64\pm 0.36$	$89.65\pm 0.59$	$0.60\pm 0.015$	$95.02\pm 0.40$	$84.09\pm 0.26$
F3Net [20]	$87.99\pm 1.45$	$85.14\pm 2.23$	$91.08\pm 0.17$	$0.58\pm 0.025$	$93.51\pm 0.63$	$86.35\pm 2.03$
DANet [28]	$88.94\pm 0.29$	$85.48\pm 2.74$	$90.50\pm 0.53$	$0.57\pm 0.015$	$94.11\pm 0.91$	$86.90\pm 1.33$
ACFNet [39]	$83.25\pm 0.18$	$83.88\pm 0.10$	$83.06\pm 0.25$	$0.34\pm 0.001$	$85.21\pm 0.15$	$90.62\pm 0.05$
CE-Net [38]	$81.49\pm 0.75$	$84.21\pm 0.85$	$84.18\pm 0.34$	$\mathbf{0}.\mathbf{30}\pm \mathbf{0}.\mathbf{005}$	$85.06\pm 0.26$	$92.00\pm 0.10$
CPFNet [40]	$85.19\pm 0.14$	$84.66\pm 1.32$	$85.22\pm 1.05$	$0.31\pm 0.002$	$86.38\pm 0.07$	$\mathbf{92}.\mathbf{09}\pm \mathbf{0}.\mathbf{20}$
ResUNet_C${}_2$F(Ours)	$\mathbf{89}.\mathbf{93}\pm \mathbf{0}.\mathbf{09}$	$\mathbf{90}.\mathbf{29}\pm \mathbf{0}.\mathbf{66}$	$\mathbf{91}.\mathbf{91}\pm \mathbf{0}.\mathbf{97}$	$0.52\pm 0.01$	$95.69\pm 0.10$	$86.75\pm 0.07$