Table 3. Parameter settings of each layer of the network (corresponding to the fur dataset).

Layer type	Convolution kernel size/pooling domain size	Number of kernels	Step size	Activation function / Pooling method
Input	-	-	-	-
Conv1	5×5	32	1	Leaky ReLU
Pool1	2×2	-	2	Max pooling
Conv1	3×3	32	1	ReLU
Pool1	2×2	-	2	Max pooling
Conv1	3×3	64	1	ReLU
Pool1	2×2	-	2	Max pooling
FC	-	-	-	-
Output	-	-	-	-