Table 2 Accuracy of cross-validation in baseline models and the proposed model.

From: Assessing central serous chorioretinopathy with deep learning and multiple optical coherence tomography images

 

Onefold

Twofold

Threefold

Fourfold

Fivefold

Average

3D-CNN

80%

82%

79%

73%

64%

75.6%

CNN-LSTM

80%

79%

86%

85%

78%

83%

VGG19 + XGB

92%

89%

91%

87%

86%

89%

VGG19 + SVM

92%

93%

90%

95%

80%

90%

VGG19 + Logistic Regression

94%

95%

90%

93%

88%

92%

ResNet-50 + Logistic Regression

91%

95%

97%

98%

90%

94.2%

  1. Significant values are in bold.
  2. 3D-CNN three-dimensional convolutional neural network, CNN-LSTM convolutional neural network–long short-term memory model, VGG19 VGG architecture with 19 layers and our custom fully connected layers, ResNet-50 ResNet architecture with 50 layers, XGB XGBoost Classifier, SVM support vector machine.
Back to article page