Single Camera based Real Time Framework for Automated Fall Detection

To identify fall of a person is the most significant issue for caring of elderly who is living alone in the home. In this paper, we propose a real-time automated fall detection and recognition framework using Support Vector Machine and vision techniques. We first capture the person using single camera and then apply Gaussian Mixture Model (GMM) method to detect the person from the video. After some preprocessing, we represent the actual human body by detecting the list of contours. After that, ellipse and rectangle shapes are fitted on these contours. Now, different important features are extracted that are useful for detection of fall activity and fed to Support Vector Machine (SVM) for the classification of fall event of a person. For the experiment, we take UR Fall Dataset that contains different activity of a person and it is available publicly. We compare our propose framework with other proposed methods and achieves better recognition accuracy.