Sensor Fusion Basketball Shooting Posture Recognition System Based on CNN

Table 1

Parameter settings of the visual geometry group network 16 (VGG16) deep learning model.

Parameters	Value
CNN layer weight initialization	Kaiming
Full connection layer weight initialization	Kaiming
Data standardization	-Score
Optimizer	Adam
Initial learning rate	1E-3
Loss function	Cross entropy
Network layer activation function	Rectified linear unit (ReLU)
Pooling method	Maximum pooling
Training rounds	20
Dropout rate	0.5
Batch size	200
Padding	[1]