Human posture recognition based on skeleton data
Kan Chen, Qiong Wang · 2015
Human posture recognition is an important research area in computer vision and it has broad application prospects in many fields, such as intelligent monitoring, human-computer interaction. But the most researches are based on the RGB image which is lack of efficiency and practicability with the effect of light and environment noise. With the skeleton data provided by Microsoft Kinect, we propose an effective and convenient way to recognize human posture. Using the preprocessing datasets of the human skeleton data, we do the comparative experiment of three classification methods with no influence of light and environmental noise. The experiment results show the efficiency of our proposed features.