Table 1.

Proposed deep CNN.

Deep CNN	Type of layer	Activation layer	Training parameter	Overall trainers
226 × 226 × 3	I/P	226 × 226 × 3	∗	—
1285 × 5 × 3	Convolution	58 × 58 × 126	Weights and bias	13,954
Rectified unit	Rectified unit	58 × 58 × 126	∗	—
Cross normalization	Normalization	58 × 58 × 126	∗	—
Maximum pooling	Max pooling	26 × 26 × 126	∗	—
Fully connected	FC	1 × 1 × 516	Weight	3,146,245
35% dropout	Dropout	1 × 1 × 514	∗	—
Fully connected layer	FC	1 × 1 × 6	Weight	2054
Softmax	Softmax	1 × 1 × 6	∗	—
O/P	Categorization	∗	∗	—