Table 1. Parameters settings for DL models utilized in the study.
Conv, convolutional layer; fc, fully connected layer; SGD, stochastic gradient descent.
| Parameters | Inception-V3 | Cascaded model | ResNet50 |
|---|---|---|---|
| Input image size | 117×117×3 | 117×117×3 | 224×224×3 |
| Input kernel size | - | - | - |
| Number of layers | 42 | 40 | 152 |
| First conv layer feature maps | 55×55×4 | 53×53×4 | 55×55×4 |
| First conv layer kernel size | 5 | 3 | 5 |
| First conv layer stride | 3×3 | 3×3 | 3×3 |
| Next few conv layer feature maps | 24×24×28 | 21×21×24 | 27×27×64 |
| No. fully connected layer (fc) | fc800 | fc800 | fc800 |
| No. of parameters in million | 23.2 | 22.8 | 25.6 |
| Batch size | 10,000 patches | 10,000 patches | 10,000 patches |
| Learning rate | 0.001 | 0.001 | 0.001 |
| Optimizer | SGD | SGD | SGD |