Table 2.
The structure and parameter settings of RV-CNN.
| Layer Name | Main Parameters | ||||
|---|---|---|---|---|---|
| Normalization Layer | Batch Normalization | ||||
| Convolution Layer | Conv. Num | Conv. Parameters | Activation | Pooling | Dropout |
| Conv. 1 | Kernel size: 5 Input num: 1 Output num: 16 |
ReLU | MaxPooling (2 × 2) | - | |
| Conv. 2 | Kernel size: 7 Input num: 16 Output num: 32 |
ReLU | MaxPooling (2 × 2) | - | |
| Conv. 3 | Kernel size: 9 Input num: 32 Output num: 64 |
ReLU | MaxPooling (2 × 2) | 50% | |
| Global Pooling Layer | Adaptive averaging pooling processing | ||||
| Fully Connected Layer | Input num: 64 Output num: 4 |
||||
| Output Layer | Softmax | ||||