TABLE VI.
Class-specific statistics on the performance of ROI-level classification. The number of true positives (TP), false positives (FP), false negatives (FN), and true negatives (TN) are given. Precision, recall (also known as true positive rate and sensitivity), false positive rate (FPR), specificity (also known as true negative rate), and F-measure are also shown.
| (a)
Penultimate-Weighted | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Class | TP | FP | FN | TN | Precision | Recall/Sensitivity | FPR | Specificity | F-measure |
| Benign | 25 | 15 | 12 | 165 | 0.6250 | 0.6757 | 0.0833 | 0.9167 | 0.6494 |
| ADH | 50 | 11 | 31 | 125 | 0.8197 | 0.6173 | 0.0809 | 0.9191 | 0.7042 |
| DCIS | 69 | 28 | 11 | 109 | 0.7113 | 0.8625 | 0.2044 | 0.7956 | 0.7797 |
| INV | 12 | 7 | 7 | 191 | 0.6316 | 0.6316 | 0.0354 | 0.9646 | 0.6316 |
| (b)
Hypercolumn-Weighted | |||||||||
| Class | TP | FP | FN | TN | Precision | Recall/Sensitivity | FPR | Specificity | F-measure |
| Benign | 29 | 21 | 8 | 159 | 0.58 | 0.7838 | 0.1167 | 0.8833 | 0.6667 |
| ADH | 44 | 8 | 37 | 128 | 0.8462 | 0.5432 | 0.0588 | 0.9412 | 0.6617 |
| DCIS | 62 | 21 | 18 | 116 | 0.747 | 0.775 | 0.1533 | 0.8467 | 0.7607 |
| INV | 15 | 17 | 4 | 181 | 0.4688 | 0.7895 | 0.0859 | 0.9141 | 0.5882 |