Table 11 Effect of different components.

From: Bladder lesion detection using EfficientNet and hybrid attention transformer through attention transformation

Components

Accuracy (%)

Precision (%)

F1-score (%)

CNN + ViT + Attention transformation

97.73

97.21

96.86

CNN

94.32

94.08

93.61

ViT + MHSA

95.12

94.85

94.59

ViT + transformation

96.48

96.11

95.81

Back to article page