Table 2. Architecture of the proposed CNN model for binary classification (one-vs-all).

Layer Name	Activation Size
Image Input	224 × 224× 3
Conv1 (7 × 7, 64) Stride 2, BN, Relu, MAP-2	112 × 112× 64
Conv2a (3 × 3, 64) Stride 1, BN, Relu	56 × 56× 64
Conv2b (3 × 3, 64) Stride 1, BN, Relu	56 × 56× 64
Conv3a (3 × 3, 128) Stride 2, BN, Relu	28 × 28× 128
Conv3b (3 × 3, 128) Stride 1, BN, Relu	28 × 28 × 128
Conv4a (3 × 3, 256) Stride 1, Relu, GAP	1 × 1 × 256
Fully Connected, SoftMax	2