Attentive 3D-Ghost Module for Dynamic Hand Gesture Recognition with Positive Knowledge Transfer

Table 7

Computational complexity comparison of the I3D network and its improved version with the 3D-Ghost module on VIVA dataset.

Model	Weights (M)	FLOPs (G)	Parameters (M)
I3D [6]	48	111.42	12.31
I3D + 3D-Ghost	45	108.16	11.23

In the experiment results, the best recognition accuracy is in bold.