Kinect based body posture detection and recognition system

A multi-class human posture detection and recognition algorithm using Kinect based geometric features is presented. The three dimensional skeletal data from the Kinect is converted to a set of angular features. The postures are classified using a support vector machines classifier with polynomial kernel. Detection of posture is done by thresholding the posture probability. The algorithm provided a recognition accuracy of 95.78% when tested using a 10 class dataset containing 6000 posture samples. The precision and recall rates of the detection system are 100% and 98.54% respectively.

[1]  Seong-Whan Lee,et al.  Gesture Spotting and Recognition for Human–Robot Interaction , 2007, IEEE Transactions on Robotics.

[2]  Janusz Konrad,et al.  A gesture-driven computer interface using Kinect , 2012, 2012 IEEE Southwest Symposium on Image Analysis and Interpretation.

[3]  Bart Selman,et al.  Human Activity Detection from RGBD Images , 2011, Plan, Activity, and Intent Recognition.

[4]  Zhengyou Zhang,et al.  Microsoft Kinect Sensor and Its Effect , 2012, IEEE Multim..

[5]  J. Montepare,et al.  The Use of Body Movements and Gestures as Cues to Emotions in Younger and Older Adults , 1999 .

[6]  Chih-Jen Lin,et al.  Probability Estimates for Multi-class Classification by Pairwise Coupling , 2003, J. Mach. Learn. Res..