Table 4 The experimental results of varying the number of blocks in each group of graph convolutional feature extraction blocks
From: R-GNN: recurrent graph neural networks for font classification of oracle bone inscriptions
Num of blocks | #Params | Top-1 | Precision | Recall | F1-score |
|---|---|---|---|---|---|
[1, 2, 4] | 21.14M | 0.873 | 0.882 | 0.854 | 0.868 |
[2, 4, 1] | 21.05M | 0.870 | 0.862 | 0.856 | 0.859 |
[4, 2, 1] | 21.14M | 0.868 | 0.843 | 0.850 | 0.847 |
[2, 2, 2] | 20.62M | 0.882 | 0.881 | 0.879 | 0.880 |