Table 3.
The modified ResNet18 Architecture for AD Classification.
| Layer Type | No of Parameters |
|---|---|
| Conv2d-1 | 9408 |
| BatchNorm2d-2 | 128 |
| Conv2d-5 | 36,864 |
| BatchNorm2d-6 | 128 |
| Conv2d-8 | 36,864 |
| BatchNorm2d-9 | 128 |
| Conv2d-12 | 36,864 |
| BatchNorm2d-13 | 128 |
| Conv2d-15 | 36,864 |
| BatchNorm2d-16 | 128 |
| Conv2d-19 | 73,728 |
| BatchNorm2d-20 | 256 |
| Conv2d-22 | 147,456 |
| BatchNorm2d-23 | 256 |
| Conv2d-24 | 8192 |
| BatchNorm2d-25 | 256 |
| Conv2d-28 | 147,456 |
| BatchNorm2d-29 | 256 |
| Conv2d-31 | 147,456 |
| BatchNorm2d-32 | 256 |
| Conv2d-35 | 294,912 |
| BatchNorm2d-36 | 512 |
| Conv2d-38 | 589,824 |
| BatchNorm2d-39 | 512 |
| Conv2d-40 | 32,768 |
| BatchNorm2d-41 | 512 |
| Conv2d-44 | 589,824 |
| BatchNorm2d-45 | 512 |
| Conv2d-47 | 589,824 |
| BatchNorm2d-48 | 512 |
| Conv2d-51 | 1,179,648 |
| BatchNorm2d-52 | 1024 |
| Conv2d-54 | 2,359,296 |
| BatchNorm2d-55 | 1024 |
| Conv2d-56 | 131,072 |
| BatchNorm2d-57 | 1024 |
| Conv2d-60 | 2,359,296 |
| BatchNorm2d-61 | 1024 |
| Conv2d-63 | 2,359,296 |
| BatchNorm2d-64 | 1024 |
| Linear-68 | 131,328 |
| Dropout-70 | 0 |
| Linear-71 | 1542 |
| LogSoftmax-72 | 0 |