Table 5 Ablation study of the proposed model components.

From: Improved pulmonary embolism detection in CT pulmonary angiogram scans with hybrid vision transformers and deep learning techniques

Model Variant

ACC(%)

Pre(%)

Rec(%)

F1-score(%)

MCC

Using ResNet50 only

79.62

72.03

69.56

74.70

0.68

Using DenseNet121 only

85.03

81.28

79.35

83.46

0.82

Using Swin Transformer only

93.75

90.22

91.06

93.18

0.81

Without Sobel filtering

94.00

91.55

93.20

92.11

0.82

Without DWT

94.93

92.11

94.40

93.14

0.85

Without AE

95.41

93.80

94.92

94.05

0.87

The proposed method

97.80

98.71

96.33

96.81

0.95

Back to article page