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

Table 1

3-fold cross-subject average accuracy (%) of different hand gesture methods on the SKIG dataset.

Method	Testing modality
	RGB	Depth
RGGP [27]	84.60	76.10
Depth context [31]	—	95.37
MRNN [32]	91.60	95.90
I3D [6]	95.74	96.58
C3D + CLSTM [15]	95.93	98.70
MTUT [5]	95.93	96.85
Ours	97.87	98.70

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