3D Action Recognition Using EigenJoints

Our Project in brief..

      The aim of this project is to recognize human actions using 3D skeleton joints recovered from 3D depth data. This 3D depth data is captured using RGB-D cameras like Microsoft Kinect.

We are using the MSR Action3D dataset which has 20 action classes and provides the required 3D joint coordinates.

The implementation is based on the algorithm proposed in the paper
"3D action recognition using eigenjoints" by Xiaodong Yang and Yingli Tian

Important Links