Table 5 Comparative evaluation between CNN and capsule networks.

From: A federated learning-based privacy-preserving image processing framework for brain tumor detection from CT scans

Model

Accuracy (%)

Precision (%)

Sensitivity (%)

Dice coefficient (%)

F1-Score (%)

Remarks

CNN

94.23

92.10

93.88

89.34

92.98

Lacks spatial sensitivity; struggles with overlapping structures

CapsuleNet

96.85

94.91

96.12

93.12

95.51

Better spatial awareness and orientation sensitivity

Aniso-ResCapHGBO-Net

99.07

98.54

99.82

96.81

99.18

Combines CapsNet spatial insight with ResNet’s depth and HGBO optimization

Back to article page