Research Article
Breast Cancer Characterization Based on Image Classification of Tissue Sections Visualized under Low Magnification
Table 2
Partial and overall classification accuracies achieved by employing the PNN classifier and the ECV method.
| Trials | Grade I% | Grade II% | Grade III% | Overall accuracy% (no. features) |
| 1 | 100 | 83.3 | 87.5 | 90.0 (3) | 2 | 100 | 83.3 | 75.0 | 85.0 (2) | 3 | 100 | 100.0 | 87.5 | 95.0 (2) | 4 | 100 | 83.3 | 87.5 | 90.0 (3) | 5 | 100 | 66.7 | 87.5 | 85.0 (3) | 6 | 100 | 50.0 | 100 | 85.0 (3) | 7 | 100 | 100 | 87.5 | 95.0 (3) | 8 | 100 | 100 | 75.0 | 90.0 (2) | 9 | 100 | 83.3 | 75.0 | 85.0 (3) | 10 | 100 | 83.3 | 100 | 95.0 (3) |
| Mean ± std | 100 ± 0 | 83.3 ± 15.7 | 86.3 ± 9.2 | 89.5 ± 4.4 |
|
|
std: standard deviation.
|