Table 1.
Performance results of image classifiers on test set.
| VGG-19 | VGG-19 | SVM | SVM | |
|---|---|---|---|---|
| (CT 1) | (CT 2) | (CT 1) | (CT 2) | |
| Accuracy | 0.783 | 0.648 | 0.728 | 0.612 |
| AUC | 0.824 | 0.78 | 0.734 | 0.682 |
| F1- Score* | 0.757 | 0.616 | 0.727 | 0.575 |
| Percentage Accuracies | ||||
| 80% Percent | 0.716 | 0.648 | 0.696 | 0.596 |
| 50% Percent | 0.684 | 0.564 | 0.704 | 0.592 |
| 30% Percent | 0.604 | 0.436 | 0.608 | 0.476 |
| 15% Percent | 0.524 | 0.328 | 0.576 | 0.376 |
| Time Cut Off Accuracies | ||||
| 15 s | 0.776 | 0.64 | 0.756 | 0.616 |
| 10 s | 0.728 | 0.604 | 0.696 | 0.620 |
| 5 s | 0.684 | 0.544 | 0.652 | 0.524 |
| 2 s | 0.608 | 0.32 | 0.592 | 0.384 |
* For Classification Task 2, weighted average metrics are reported. Total number of test observations = 250. The best result of each trial is highlighted in bold.