Skip to main content

Table 1 Results (mean ± standard deviation) of ten CNN model architectures after ten averages (from high to low accuracy)

From: Automatic ovarian tumors recognition system based on ensemble convolutional neural network with ultrasound imaging

CNN models

ResNet-50

ResNet-101

ResNet-18

DenseNet-201

Inception-V3

ACC (%)

90.51 ± 4.36

90.15 ± 3.03

89.90 ± 3.30

89.79 ± 2.84

89.79 ± 3.70

SE (%)

89.02 ± 5.44

89.65 ± 3.86

89.05 ± 5.27

89.77 ± 4.16

89.50 ± 4.11

SP (%)

92.00 ± 5.95

90.67 ± 3.86

90.77 ± 4.66

89.79 ± 4.10

90.06 ± 4.99

CNN models

Darknet-53

ShuffleNet

Xception

Darknet-19

MobileNet-v2

ACC (%)

89.53 ± 1.80

89.40 ± 3.78

88.79 ± 1.21

88.28 ± 1.90

87.41 ± 1.83

SE (%)

87.85 ± 3.99

87.43 ± 4.41

88.64 ± 2.95

88.05 ± 3.53

86.85 ± 1.98

SP (%)

91.20 ± 3.16

91.35 ± 3.89

88.93 ± 3.47

88.46 ± 3.36

87.99 ± 2.88