People Tracking in a Smart Room Using Kinect Sensor
Heba A. Salim, Musaab Alaziz, Turki Younis Abdalla · 2021
In this paper, a new method for people tracking in a smart room using a Kinect sensor is proposed. The approach is based on the skeleton data with the (X, Y, Z) coordinate values of each joint in the human body which is provided by the Kinect sensor. For data classification, the Support Vector Machine (SVM) technique is used. To achieve this goal 14 movement classes are defined. Experiments were conducted on 12 subjects each one performs 14 movements in each experiment, the training dataset is created manually by capturing the subject movements during all experiments. The result of these we get after training the SVM model shows that the average accuracy is 90.2%.