TABLE II. Comparison of Performances With Other State-of-The-Art Networks on COVID Lesion Segmentation on MosMedData.

Networks	Sensitivity (%)	Precision (%)	Dice (%)	IoU (%)
FCN [27]	78.8 0.23	58.9 0.16	35.6 0.36	29.3 0.45
Unet [16]	94.3 0.34	74.4 0.32	50.5 0.26	40.3 0.23
Vnet [28]	84.5 0.42	64.6 0.54	40.2 0.33	36.4 0.26
Unet++ [17]	78.1 0.15	65.1 0.25	37.2 0.27	33.3 0.32
CPF-Net [29]	82.4 0.25	71.3 0.29	48.9 0.21	37.6 0.38
COPLE-Net [14]	85.5 0.18	73.1 0.20	51.1 0.21	41.2 0.38
Mini-SegNet [13]	81.5 0.25	69.1 0.19	43.7 0.23	35.2 0.38
Inf-Net [12]	92.8 0.27	76.9 0.34	51.8 0.31	41.6 0.27
TA-SegNet (Prop.)	99.6 0.09	84.8 0.26	62.3 0.18	51.7 0.29