Table 3 Performance metrics in the testing set for the slices evaluation using the best model found by the Hyperparameter tuning.
From: Contrast phase recognition in liver computer tomography using deep learning
Precision | Recall | F1-score | AUC | N | |
|---|---|---|---|---|---|
Unenhanced | 0.99 | 0.99 | 0.99 | 1 | 10,757 |
Arterial | 0.98 | 0.98 | 0.98 | 1 | 10,870 |
Portal | 0.92 | 0.89 | 0.89 | 0.99 | 11,907 |
Delayed | 0.89 | 0.93 | 0.91 | 0.99 | 10,803 |
Average | 0.95 | 0.95 | 0.95 | 0.99 | – |
