Table 4 Comparison of accuracy using different dimensional fusion of structured data and pathological image
From: Richer fusion network for breast cancer classification based on multimodal data
Method | Accuracy (validation) (%) | Accuracy (test) (%) |
|---|---|---|
Structured data only (29D) | 90.3 | 78.5 |
Pathological image only (VGG16) | 84.5 | 83.6 |
Structured data (29D) + Pathological image (FC-29D) | 88.2 | 87.9 |
Structured data (29D) + Pathological image (FC-1280D) | 88.6 | 85.2 |
Structured data (29D*10) + Pathological image (FC-1280D) | 89.2 | 88.6 |
Structured data (29D*20) + Pathological image (FC-1280D) | 93.3 | 91.1 |
Structured data (29D*30) + Pathological image (FC-1280D) | 92.5 | 89.2 |
Structured data (29D*40) + Pathological image (FC-1280D) | 89.5 | 88.7 |