Visual feature-based violent video detection
Xin Ji, Ou Wu, Chenlong Wang, Jinfeng Yang · 2014
In this paper, a novel violent video detection scheme is presented based on visual views. The violent segment is detected by using motion angle, motion intensity, shot & explosion and blood analysis. In addition, in order to improve the compute efficiency and time savings Hidden Information SVM is used. Experimental result shows that the proposed method is effective in violent video detection.