Table 7 Result of Feature Connected Layer Output Combination (FCLOMB + IRD) for RSNA dataset
From: Multi Region-Based Feature Connected Layer (RB-FCL) of deep learning models for bone age assessment
Regressor | InceptionV3-RB-FCL + IRD | DenseNet121-RB-FCL + IRD | InceptionResNetV2-RB-FCL + IRD | FCLOMB + IRD | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
MAE | RMSE | MAPE% | MAE | RMSE | MAPE% | MAE | RMSE | MAPE% | MAE | RMSE | MAPE% | |
LR | 7.88 | 10.48 | 9.524 | 7.36 | 9.763 | 8.609 | 7.79 | 10.38 | 9.734 | 7.32 | 9.754 | 8.627 |
KNN-R | 7.78 | 10.4 | 8.74 | 7.2 | 9.722 | 7.92 | 7.71 | 10.37 | 9.031 | 7.11 | 9.583 | 7.912 |
SVR | 8.04 | 11 | 11.57 | 8.22 | 12.29 | 14.72 | 8.01 | 11.02 | 11.93 | 8.17 | 12.53 | 15.13 |
RF-R | 7.55 | 10.08 | 8.153 | 7.07 | 9.497 | 7.807 | 7.53 | 10.25 | 8.931 | 6.97 | 9.346 | 8.128 |
DT-R | 10.5 | 14.66 | 10.87 | 9.63 | 13.61 | 9.819 | 10.3 | 14.68 | 11.06 | 9.29 | 12.94 | 9.515 |
ADB-R | 11.8 | 15.45 | 13.55 | 11.1 | 14.15 | 12.72 | 11.8 | 15.26 | 13.23 | 11.1 | 14.2 | 13.59 |
GB-R | 7.97 | 10.61 | 9.581 | 7.61 | 10.07 | 8.485 | 7.74 | 10.52 | 9.587 | 7.35 | 9.849 | 8.703 |