Open Access
Table 2.
Summary of performance metrics for different training scenarios (see learning curves in Fig. 8).
| Metric/Scenario | LS | HST | Stacked images | Merged branches |
|---|---|---|---|---|
| Epochs for Training | 304 | 327 | 307 | 321 |
| Learning Curve Characteristics | Fast initial learning with moderate stability post-training | Slower convergence due to noise, artifacts, and higher resolution | Swift decrease, with minimal fluctuations after 100 epochs; stable with small generalization gap | Gradual decrease in validation loss; stable with slower convergence due to added complexity |
| TPR at FPR of 10−4 | 0.45 | 0.41 | 0.51 | 0.55 |
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.