Table 1
Evaluation metrics for the intensity model.
| Loss function | Learning rate | RMSE (PPF) | MAE (PPF) | COS | CC | SSIM |
|---|---|---|---|---|---|---|
| MSE | 0.1 | 1.124 | 0.872 | 0.997 | 0.751 | 0.781 |
| MAE | 0.1 | 1.076 | 0.821 | 0.997 | 0.772 | 0.786 |
| MASE | 0.01 | 1.042 | 0.79 | 0.997 | 0.784 | 0.791 |
| LCOS | 0.01 | 1.062 | 0.809 | 0.997 | 0.776 | 0.790 |
Notes. The four loss functions are MS E for the mean square error, MAE for the mean absolute error, MAS E for a combination of MS E and MAE, and LCOS for a customized loss function. The evaluation metrics contain the RMS E (root mean square error), MAE, COS (average cosine of the angle between the reconstructed velocity field and the ground truth), CC (correlation coefficient), and S S IM (structural similarity). See the main text for the definitions and formulas of the loss functions and the evaluation metrics. Values in red are the best ones of each evaluation metric.
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.