Table 2. Confusion matrices for the benignity and malignancy classification in test sets A and B.
Ground truth | |||||
---|---|---|---|---|---|
Test A | Test B | ||||
Class | benign | malignant | benign | malignant | |
Prediction | benign | 32a | 0b | 4 | 1 |
malignant | 1c | 27d | 0 | 15 | |
33 | 27 | 4 | 16 | ||
Samples | 60 | 20 |
Notes.
For computation of class-wise tissue classification performance we use the following values:
True positives (TP).
False positives (FP).
False negatives (FN).
True negatives (TN) are not considered in the class-wise metrics, as we report precision and recall for both classes separately in Table 3.