Table 4 Comparisons with existing methods.

From: Semantic segmentation of COVID-19 lesions with a multiscale dilated convolutional network

Network

Number of parameters (M)

Inference time (ms)

Dice (%)

Sen. (%)

Spec. (%)

mIoU (%)

FCN27

3.8

156

67.9

70.7

85.3

62.5

U-Net24

2.4

100

73.3

76.6

98.6

67.1

SegNet28

3.1

89.2

70.5

77.8

94.8

65.6

Dilated-1030

6.5

78

75.2

78.1

92.3

70.3

MSDC-Net (ours)

5.7

63

82.4

81.1

97.7

78.2

Back to article page