Using a Selective Ensemble Support Vector Machine to Fuse Multimodal Features for Human Action Recognition

Table 4

Recognition rate using the single modal feature and multimodal features.

Dataset	Descriptor	Precision (%)
G3D	RGB-HOG	83.7
	DMM-LBP	83.2
	HJF	83.7
	Mixed features	91.7
CAD60	RGB-HOG	85.3
	DMM-LBP	86.0
	HJF	88.6
	Mixed features	91.8