Table 1:

Schematic of the proposed modified Inception V4 CNN. The input size is given in each row, and the output size is the input size of the next row. All convolutions were performed with sigmoid activation and batch normalization.

Layer	Kernel size / remarks	Input Size
Conv	3×3 / ‘valid’	101×101×3
Conv	3×3 / ‘valid’	98×98×32
Max Pool	2×2 / stride=2, ‘valid’	96×96×64
Conv	3×3 / stride=2, ‘valid’	48×48×64
4× Inception A Block	1×1, 3×3 / ‘same’	23×23×80
1× Reduction A Block	1×1, 3×3 / ‘valid’	23×23×384
7× Inception B Block	1×1, 1×7, 7×1, 3×3 / ‘same’	11×11×1024
1× Reduction B Block	1×1, 1×7, 7×1, 3×3 / ‘valid’	11×11×1024
3× Inception C Block	1×1, 1×3, 3×1, 3×3 / ‘same’	5×5×1536
Average Pool	5×5 / ‘valid’	5×5×1536
Linear	Logits	1 × 1536
Softmax	Classifier	1 × 2