We use Sensitivity and Specificity to evaluate the performance
of classification algorithms. Sensitivity (or True
Positive Rate) measures the proportion of images, which actually are UML CDs, are correctly identified as UML CDs.
Specificity (or True Negative Rate) denotes the proportion of
actual non-UML CD images that are correctly classified as
non-UML CDs. In other words, while specificity represents
the ability of excluding non-UML CD images, sensitivity
represents the ability of including UML CD images. The two
metrics are calculated from the confusion matrix as below: