Table 2. Test result of the trained models with unseen test dataset

Decoder	Encoder	Dice	IoU	Accuracy	Recall	Precision
FCN8	VGG16	97.90 ± 0.25^ab	95.97 ± 0.45^ab	99.37 ± 0.07^a	96.81 ± 0.44^ab	99.06 ± 0.12^ab
	ResNet50	98.31 ± 0.44^a	96.75 ± 0.81^a	99.49 ± 0.13^a	97.50 ± 0.76^a	99.17 ± 0.09^a
	MobileNet	97.14 ± 0.16^abc	94.57 ± 0.29^abc	99.15 ± 0.04^a	95.52 ± 0.37^abc	98.91 ± 0.17^ab
FCN16	VGG16	96.91 ± 0.41^abc	94.17 ± 0.72^abc	99.08 ± 0.12^a	95.45 ± 0.63^abc	98.48 ± 0.28^abc
	ResNet50	96.38 ± 1.20^bc	93.39 ± 2.02^bc	98.96 ± 0.32^a	94.46 ± 1.92^abcd	98.67 ± 0.22^ab
	MobileNet	94.44 ± 0.19^de	89.93 ± 0.31^de	98.39 ± 0.05^a	91.90 ± 0.53^de	97.40 ± 0.49^c
FCN32	VGG16	89.70 ± 0.21^f	81.77 ± 0.25^f	93.96 ± 1.07^c	88.73 ± 1.20^f	91.13 ± 1.30^e
	ResNet50	94.00 ± 0.51^e	89.23 ± 0.83^e	98.29 ± 0.13^ab	90.82 ± 0.77^ef	97.81 ± 0.36^bc
	MobileNet	88.51 ± 0.86^f	81.11 ± 1.16^f	96.96 ± 0.18^b	83.61 ± 1.18^g	95.43 ± 0.43^d
SegNet	VGG16	96.45 ± 0.75^bc	93.40 ± 1.29^bc	98.97 ± 0.20^a	94.04 ± 1.21^bcd	99.24 ± 0.17^a
	ResNet50	98.04 ± 0.54^ab	96.25 ± 1.01^ab	98.05 ± 1.06^ab	96.85 ± 1.10^ab	99.36 ± 0.13^a
	MobileNet	95.73 ± 0.33^cd	92.09 ± 0.56^cd	98.77 ± 0.09^a	92.83 ± 0.61^cde	99.12 ± 0.09^ab
U-Net	VGG16	98.34 ± 0.49^a	96.81 ± 0.90^a	99.47 ± 0.81^a	97.38 ± 0.83^a	99.37 ± 0.13^a
	ResNet50	98.26 ± 0.42^a	96.66 ± 0.78^a	99.18 ± 0.27^a	97.11 ± 0.82^a	99.52 ± 0.09^a
	MobileNet	98.35 ± 0.24^a	96.80 ± 0.45^a	99.50 ± 0.07^a	97.20 ± 0.47^a	99.57 ± 0.07^a

The values are represented as mean ± standard error mean (p < 0.05).
Dice, dice coefficient; IoU, intersection of union.