Research Article
Feature Fusion Based on Convolutional Neural Network for Breast Cancer Auxiliary Diagnosis
Table 2
Comparison of the results of various indicators under different magnifications.
| Enlarge size | Model | Accuracy (%) | (%) | (%) | (%) |
| 40x | VGG16 | 94.39 | 92.99 | 94.22 | 93.56 | InceptionV3 | 92.20 | 97.73 | 89.85 | 90.71 | ResNet50 | 84.15 | 81.76 | 85.70 | 82.79 | VIRNets (ours) | 99.02 | 98.86 | 98.97 | 98.90 |
| 100X | VGG16 | 86.19 | 83.51 | 86.61 | 84.63 | InceptionV3 | 96.67 | 95.78 | 97.93 | 96.74 | ResNet50 | 90.95 | 88.64 | 91.75 | 89.86 | VIRNets (ours) | 97.62 | 96.43 | 98.28 | 97.27 |
| 200X | VGG16 | 83.66 | 80.85 | 81.62 | 81.21 | InceptionV3 | 97.03 | 98.12 | 97.26 | 97.67 | ResNet50 | 91.58 | 91.43 | 88.68 | 89.88 | VIRNets (ours) | 98.02 | 97.02 | 98.56 | 97.73 |
| 400X | VGG16 | 86.41 | 84.84 | 83.9 | 84.34 | InceptionV3 | 95.56 | 94.27 | 96.34 | 95.17 | ResNet50 | 83.15 | 80.73 | 82.77 | 81.50 | VIRNets (ours) | 98.37 | 97.95 | 98.36 | 98.15 |
|
|