Table 2.
The configuration of the proposed model with the used parameters
| Layer type | Number of filters | Kernel size | Output size | Number of parameter |
|---|---|---|---|---|
| Conv2D | 32 | 3 × 3 | (254, 254, 32) | 320 |
| Conv2D | 10 | 1 × 1 | (254, 254, 10) | 330 |
| AveragePooling2D | – | 2 × 2 | (127, 127, 10) | 0 |
| Conv2D | 10 | 1 × 1 | (127, 127, 10) | 110 |
| Conv2D | 32 | 3 × 3 | (125, 125, 32) | 2912 |
| Conv2D | 10 | 1 × 1 | (125, 125, 10) | 330 |
| AveragePooling2D | - | 2 × 2 | (62, 62, 10) | 0 |
| Conv2D | 10 | 1 × 1 | (62, 62, 10) | 110 |
| Conv2D | 32 | 3 × 3 | (60, 60, 32) | 2912 |
| Conv2D | 10 | 1 × 1 | (60, 60, 10) | 330 |
| AveragePooling2D | – | 2 × 2 | (30, 30, 10) | 0 |
| Flatten | – | – | 9000 | 0 |
| Dense | – | – | 64 | 576,064 |
| Softmax | – | – | 3 | 195 |
| Total parameters | 583,613 | |||
| Trainable parameters | 583,613 |