Table 2 Hyperparameters for deep CNN Classifier.
Layer | Filters | Kernel | Input size | Output size | Description |
|---|---|---|---|---|---|
Convolutional 1 | 64 | 3 × 3 | 128 × 128 × 1 | 128 × 128 × 64 | Extracts low- level features from input data |
Max Pooling 1 |  | 2 × 2 | 128 × 128 × 64 | 64 × 64 × 64 | Reduces spatial dimensions, retains key features |
Convolutional 2 | 128 | 3 × 3 | 64 × 64 × 64 | 64 × 64 × 128 | Further feature extraction with increased depth |
Max Pooling 2 |  | 2 × 2 | 64 × 64 × 128 | 32 × 32 × 128 | Down sampling to reduce dimensionality |
Convolutional 3 | 256 | 3 × 3 | 32 × 32 × 128 | 32 × 32 × 256 | Advanced feature extraction from deeper layers |
Max Pooling 3 |  | 2 × 2 | 32 × 32 × 256 | 16 × 16 × 256 | Further reduction in spatial size |
Convolutional 4 | 512 | 3 × 3 | 16 × 16 × 256 | 16 × 16 × 512 | Higher-level feature extraction |
Max Pooling 4 |  | 2 × 2 | 16 × 16 × 512 | 8 × 8 × 512 | Continues reducing dimensionality |
Convolutional 5 | 512 | 3 × 3 | 8 × 8 × 512 | 8 × 8 × 512 | Continues extracting features at a fine level |
Max Pooling 5 |  | 2 × 2 | 8 × 8 × 512 | 4 × 4 × 512 | Reduces spatial dimensions for classification |
Convolutional 6 | 1024 | 3 × 3 | 4 × 4 × 512 | 4 × 4 × 1024 | Final feature extraction with deeper network |
Max P ooling 6 |  | 2 × 2 | 4 × 4 × 1024 | 2 × 2 × 1024 | Reduces dimensions to finalize feature extraction |
Dropout |  |  | 2 × 2 × 1024 | 2 × 2 × 1024 | Regularization to prevent overfitting |
Flatten |  |  | 2 × 2 × 1024 | 1 × 4096 | Flattens the feature map into a 1D vector |
Fully connected | 4096 |  | 1 × 4096 | 1 × N | Final classification layer (N =  number of classes) |
