Table 4 The effect of comprehensive attention model on Accuracy (%)
From: CRANet: a comprehensive residual attention network for intracranial aneurysm image classification
Mode structure | Accuracy |
|---|---|
SA1 + CA1 | 96.91 ± 0.1 |
SA1-2 + CA1-2 | 97.20 ± 0.2 |
SA1-3 + CA1-3 | 97.36 ± 0.1 |
SA1-4 + CA1-4 | 97.55 ± 0.3 |
SA1-5 + CA1-5 | 97.81 ± 0.2 |
CA1 + SA1 | 96.64 ± 0.4 |
CA1-2 + SA1-2 | 97.11 ± 0.1 |
CA1-3 + SA1-3 | 97.26 ± 0.3 |
CA1-4 + SA1-4 | 97.42 ± 0.4 |
CA1-5 + SA1-5 | 97.63 ± 0.1 |