Detection of Fungal Infections in Gloriosa Superba Plant Using the Convolution Neural Network Model

Table 2

PSO for CNN hyperparameters.

Variables	HPs	Range
y1	Number of first convolution kernels	[1, 300]
y2	First convolution layer kernel size	{33, 55, 77, 99}
y3	First convolution layer activation function	{Sigmoid, ReLu, Tanh}
y4	First pooling layer types	{Max-pooling, min-pooling, average-pooling}
y5	Number of second convolution kernels	[1, 300]
y6	The 2^nd convolution layer kernel size	{33, 55, 77, 99}
y7	The 2^nd convolution layer activation function	{ReLu, Sigmoid, Tanh}
y8	The 2^nd pooling layer	{max-pooling, min-pooling, average-pooling}
y9	The first FC layer neuron	[10, 800]
y10	Activation function of the first FC layer	{ReLu, Sigmoid, Tanh}
y11	First FC layer dropout	[0.1, 0.9]
y12	Number of second FC layer neurons	[10, 800]
y13	The 2^nd FC layer activation function	{Sigmoid, ReLu, Tanh}
y14	The 2^nd FC layer dropout	[0.1, 0.8]
y15	Learning rate	[0.01, 0.001, 1]