Open Access
Table 1.
Optimized hyperparameters controlling the Siamese CNN architecture.
| Parameter | Function | Value |
|---|---|---|
| Embedding dimension | Dimensionality of the learned feature space | 64 |
| Triplet loss margin | Minimum separation between positive and negative pairs in embedding space | 0.1 |
| Learning rate | Step size during optimization | 0.0001 |
| Number of epochs | Total training iterations | 100 |
| Batch size | Number of samples per model update | 32 |
| Steps per epoch | Batch iterations per epoch | 75 |
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.