Research Article
Feature Fusion Based on Convolutional Neural Network for Breast Cancer Auxiliary Diagnosis
Table 3
Comparison of the correct rate of the VIRNets model using the BreaKHis test set and different architectures.
| Structure | Accuracy under different magnifications | Average accuracy (%) | 40X (%) | 100X (%) | 200X (%) | 400X (%) |
| CSDCNN [19] | 95.9 | 96.9 | 96.7 | 94.9 | 96.1 | IRRCNN [20] | 97.95 | 97.57 | 97.32 | 97.36 | 97.55 | Bi-LSTM [21] | 96.2 | 97.2 | 97.1 | 95.4 | 96.48 | BreastNet [22] | 97.99 | 97.84 | 98.51 | 95.88 | 97.55 | Sharma et al. [23] | 97.4 | 98.6 | 97.7 | 96.8 | 97.63 | VIRNets (ours) | 99.02 | 97.62 | 98.02 | 98.37 | 98.26 |
|
|