Human action recognition based on Kinect and PSO-SVM by representing 3D skeletons as points in lie group

In this paper, we propose an effective method to recognize human actions. Combined with the relationship between 3D skeleton model of joint position and particle group optimization algorithm is used to optimize the support vector machine (PSO-SVM) and depth through the Kinect sensor to obtain human 3D skeleton model, each skeletal model with 20 joints and 19 joints, the relative geometry between various body parts provides a more meaningful description than their absolute locations, we explicitly model the relative 3D geometry between different body parts in our skeletal representation. Mathematically, rigid body rotations and translations in 3D space are members of the special Euclidean group SE(3), which is a matrix Lie group. Hence, we represent the relative geometry between a pair of body parts as a point in SE(3), We then perform classification using a combination of dynamic time warping, and particle swarm optimization on support vector machine (PSO-SVM), Experimental results on three action datasets: MSR-Action 3D, UT-Kinect, Florence 3D-Action, show that the proposed representation performs better than many existing skeletal representations. The proposed approach also outperforms various state-of-the-art skeleton-based human action recognition approaches.