An Improved Crop Disease Identification Method Based on Lightweight Convolutional Neural Network

Table 1

Parameters setting in the proposed CNN.

Parameters	Information
Activation function	ReLU
Optimizer	Adam
Convolution type	Same
Convolution kernel	3 ∗ 3, 5 ∗ 5, 7 ∗ 7
Convolution stride	1
Pooling type	Average pooling, global average pooling
Pooling stride	2
Pooling size	2 ∗ 2