Table 2.

Comparative results of U-Net with the different backbones and the batch size of 8 (bold represent the best performance).

Model Name	Backbone	Batch Size	Optimizer	Loss Function	Metrics
					Accuracy	Sensitivity	Specificity	Dice	AJI
U-Net	ResNet101	8	Adam	BCE	0.98939	0.67254	0.99584	0.68844	0.53922
	EfficientNetB7				0.98992	0.7392	0.99526	0.72448	0.57832
	DenseNet161				0.98881	0.66545	0.99521	0.66838	0.51961
	InceptionResNetV2				0.98918	0.66615	0.99553	0.67742	0.52953
	SENetResNext101				0.98812	0.74222	0.99331	0.67823	0.53138
	MobileNetV2				0.98891	0.63465	0.99589	0.65913	0.50924
	VGG19				0.98778	0.5761	0.99568	0.61238	0.46402