Command recognition by using skeletal tracking based on Kinect sensor
Çağrı Akyildiz, Şafak Saraydemi̇r · 2016
In this study, a command recognition method which works responsively to human body motion is developed by using a skeletal tracking system based on Microsoft Kinect sensor in MATLAB environment. Nowadays, image processing techniques that are used by benefiting from Kinect sensor are developing rapidly. The skeletal tracking system algorithm used in this study also works using the depth mapping ability of Kinect sensor. Firstly, the human body in sight of Kinect sensor is detected and afterwards the similarity of gestures and predefined commands in the system are compared. If the commands received from skeletal inference of user in incoming image data don't exceed the determined tolerance range, the gesture is identified by the system and presented as output. During the test stage of study; the 8 command gesture images taken from each of 9 participants are used as the data set and the correct identification of commands with high accuracy is observed.