Using a Multilearner to Fuse Multimodal Features for Human Action Recognition

Table 2

Comparison of various action recognition methods.

Dataset	Descriptor	This paper’s method (%)	Boosting (%)	Bagging (%)	SVM (%)	ANNs (%)
G3D	RGB-HOG	82	79	75	79	74
	D-STIP	81	80	79	80	80
	S-JRPF	84	81	80	80	74
Ours	93	—	—	—	—
CAD60	RGB-HOG	88	84	80	81	76
	D-STIP	87	80	79	80	80
	S-JRPF	88	81	80	80	74
	Ours	94	—	—	—	—