Table 2. Network architecture of the CNN model.

The first column depicts the different layers used consecutively in the network. The ”layer shape” column reports the shape of the convolutional kernels, the max-pooling windows and the fully connected layers. The ”output shape” column reports the variation of layer shapes at each step.

Layer name	Layer shape	Output shape
Input	–	4 × 299 × 1
Conv2D	32 × 4 × (4 × 1)	32 × 296 × 1
Max-pooling	2 × 1	32 × 148 × 1
Dropout	–	32 × 148 × 1
Conv2D	64 × 32 × (4 × 1)	64 × 145 × 1
Max-pooling	2 × 1	64 × 72 × 1
Dropout	–	64 × 72 × 1
Conv2D	128 × 64 × (4 × 1)	128 × 69 × 1
Max-pooling	2 × 1	128 × 34 × 1
Dropout	–	128 × 34 × 1
Dense	128	128
Dropout	–	128
Dense (sigmoid)	1	1