Human motion tracking and estimation of critical points of GGVF of human body for behavior recognition
Shengwen Guo, Jinshan Tang · 2010
Human motion tracking and analysis have received increasing attention for motion capture and human behavior recognition, which aim to extract useful information such as position, pose, and velocity of a moving human body from image sequences (video). This paper proposed a new human motion tracking method which searches critical points of generalized gradient vector flow (GGVF) around the skeleton of a human body. We focus on the motion analysis of critical points since it reveals the essential movement of the observed object. Experimental results demonstrated that the proposed method can extract crucial information of the dynamic body motions.