Hand Gesture Classification Based on Nonaudible Sound Using Convolutional Neural Network

Table 2

Average, minimum, and maximum classification accuracies according to data and environment.

Environment	Input data	Accuracy (%)
		Average	Minimum	Maximum
w/o noise	Raw	76.25	70.62	81.25
	STFT	87.75	85.62	92.5
w/ noise	Raw	74.18	69.37	78.12
	STFT	79.37	72.18	84.37