TABLE 6. Performance Comparison Between Different Layers of the Proposed CNN-Based Framework.

	1-layer Conv+1-layer FCO	2-layer Conv+1-layer FCO,	3-layer Conv+1-layer FCO	4-layer Conv+1-layer FCO	4-layer CNN+2-layer FCN	4-layer Conv+1-layer FC+1-layer FCO (Proposed Model)
AUC	0.689	0.719	0.922	0.947	0.956	0.969
Accuracy	0.461	0.522	0.820	0.932	0.959	0.968
F1	0.365	0.411	0.737	0.888	0.933	0.945
Sensitivity	0.425	0.459	0.780	0.917	0.949	0.960
Specificity	0.810	0.824	0.940	0.978	0.986	0.989
Precision	0.369	0.415	0.709	0.865	0.919	0.932
Recall	0.425	0.459	0.780	0.917	0.949	0.960

Conv: convolutional layer. FC: fully-connected layers. FCO: fully-connected output layer.