Table 2.
Hyperparameters used in the network
| Layer no. and type | Activation function used | Regularization used | No. of feature maps/nodes (L + R) | Kernel Size | Stride length | Pooling size | Pooling Stride length |
|---|---|---|---|---|---|---|---|
| 1 (Convolutional Layer) | Relu | L2 | 2 + 2 | (5,5) | (1,1) | (2,2) | (2,2) |
| 2 (Convolutional Layer) | Relu | L2 | 4 + 4 | (5,5) | (1,1) | – | – |
| 3 (Convolutional Layer) | Relu | L2 | 8 + 8 | (5,5) | (1,1) | – | – |
| 4 (Convolutional Layer) | Relu | L2 | 4 + 4 | (5,5) | (1,1) | – | – |
| 5 (Convolutional Layer) | Relu | L2 | 2 + 2 | (5,5) | (1,1) | – | – |
| 6 (Fully connected layer) | Sigmoid | L2 | 50 + 50 | – | – | – | – |
| 7 (Fully connected layer) | Sigmoid | L2 | 30 + 30 | – | – | – | – |
| 8(Fully connected layer) | Sigmoid | L2 | 6 + 6 | – | – | – | – |
Learning rate used with the network is 0.0001