Table 2 The mean absolute error(MAE), Mean Squared Error (MSE), Pearson correlation(PCC) and 95% confidence intervals comparison of Siamese networks for predicting reproduction skill scores under different backbone networks in the test set.The lower the MAE and MSE, the smaller the difference between the predicted value and the measured value; The higher the PCC, the higher the linear correlation between the predicted value and the measured value.
From: Predict the writer’s trait emotional intelligence from reproduced calligraphy
Evaluation Indicator | VGG16 | VGG19 | Resnet50 | Resnet101 | Densenet121 |
|---|---|---|---|---|---|
MAE | 0.390 | 0.398 | 0.529 | 0.563 | 0.596 |
MSE | 0.251 | 0.265 | 0.461 | 0.525 | 0.567 |
PCC | 0.770 | 0.755 | 0.575 | 0.498 | 0.523 |
95% CI of PCC | [0.758, 0.781] | [0.743, 0.767] | [0.556, 0.593] | [0.477, 0.519] | [0.502, 0.543] |