Table 3.
Convolutional neural network and deep feature SRC classification performance.
| Method | TPR (%) | TNR (%) | ACC (%) | AUC (%) |
|---|---|---|---|---|
| End-to-End CNN | ||||
| Googlenet | 68.81 | 52.06 | 60.44 | 64.55 |
| InceptionV3 | 56.19 | 73.71 | 64.95 | 68.96 |
| ResNet18 | 59.28 | 64.18 | 61.73 | 65.14 |
| ResNet50 | 53.09 | 72.94 | 63.02 | 67.82 |
| ResNet101 | 56.44 | 79.64 | 68.04 | 72.05 |
| Densenet201 | 62.37 | 82.22 | 72.29 | 77.51 |
| InceptionResNetV2 | 60.31 | 65.98 | 63.14 | 67.03 |
| Xception | 54.90 | 82.73 | 68.81 | 72.69 |
| DF-SRC | ||||
| Googlenet-SRC | 61.24 | 61.89 | 61.59 | 63.93 |
| InceptionV3-SRC | 66.41 | 72.52 | 69.63 | 75.23 |
| ResNet18-SRC | 67.47 | 67.21 | 64.27 | 67.47 |
| ResNet50-SRC | 62.27 | 68.13 | 65.37 | 69.16 |
| ResNet101-SRC | 62.79 | 73.21 | 68.29 | 72.45 |
| Densenet201-SRC | 65.12 | 71.13 | 68.29 | 72.30 |
| InceptionResNetV2-SRC | 67.70 | 72.75 | 70.37 | 72.58 |
| Xception-SRC | 59.43 | 70.67 | 65.37 | 67.67 |