Skip to main content

Table 2 The mean F-score, precision, recall, and specificity for each deep learning model using 90/10 data split

From: Detection of developmental dysplasia of the hip in X-ray images using deep transfer learning

Deep learning model

F1 score

Precision

Recall

Specificity

SqueezeNet

92.3%

86.2%

100.0%

90.9%

GoogLeNet

86.5%

79.6%

96.7%

84.8%

Inceptionv3

90.0%

82.1%

100.0%

88.3%

DenseNet-201

86.2%

76.6%

99.2%

83.5%

MobileNetv2

94.6%

90.1%

100.0%

93.9%

ShuffleNet

82.9%

71.6%

98.8%

79.6%

Resnet101

92.5%

86.3%

100.0%

91.3%

Resnet50

94.4%

89.8%

100.0%

93.5%

Resnet18

89.1%

81.4%

99.2%

87.4%

Xception

83.9%

76.4%

93.3%

84.8%

Inception-ResNet-v2

86.1%

76.4%

99.2%

83.5%

ShuffleNet

87.7%

78.9%

99.2%

85.7%

DarkNet-53

95.0%

90.6%

100.0%

94.3%

EfficientNet-b0

87.7%

78.3%

100.0%

85.2%