Table 3 Type identification models Performance with geometric augmentation.

From: Vision transformer and deep learning based weighted ensemble model for automated spine fracture type identification with GAN generated CT images

Ensemble models with geometric augmentation

Accuracy (%)

Recall

F1-Score

Resnet-DenseNet-Ensemble Model

57.67

0.58

0.51

VGG16-DenseNet-Ensemble Model

64.20

0.64

0.60

VGG16-Resnet-DenseNet-Ensemble Model

66.84

0.67

0.63

VGG16-Resnet-Ensemble Model

81.83

0.82

0.81

ViT

86.28

0.82

0.82

VGG16-ResNet50-ViT-Proposed Ensemble Model

89.98

0.82

0.89

Back to article page