An Action Recognition Method for Volleyball Players Using Deep Learning

Table 3

Influence of different numbers of feature maps on the recognition effect.

	Combination 1	Combination 2	Combination 3
Number of characteristic graphs Average recognition rate	The first layer is 40	The first layer is 46	The first layer is 60
	The second convolutional layer is 60	The second convolutional layer is 78	The second convolutional layer is 90
Florence3D action (%)	91.00	92.00	93.00