[Retracted] Dynamic Gesture Recognition Algorithm Based on 3D Convolutional Neural Network

Table 1

Results of different network models in EgoGesture.

Recognition methods	Input categories	Validation set of EgoGseture (%)	Test set of EgoGseture (%)
Harris3.5D	RGB + depth	35.8	36.1
HOG3D	RGB + depth	43.8	44.6
HON4D	Depth	57.2	58.3
C3D	RGB + depth	56.1	57.4
T3D	RGB + depth	62.4	63.8
R3D	RGB + depth	64.8	66.1
Proposed algorithm	RGB + depth	71.5	72.4