Table 2 . A summary of the architecture of the four pre-trained CNNs used.
| CNN architecture | Number of layers | Input size | Output size |
|---|---|---|---|
| AlexNet | 8 | 227 × 227 | 4,096 × 2 |
| GoogleNet | 22 | 224 × 224 | 1,024 × 2 |
| ResNet-18 | 18 | 224 × 224 | 512 × 2 |
| ShuffleNet | 50 | 224 × 224 | 544 × 2 |