An Elderly Fall Detection System using Depth Images

In this proposed conceptual technique, we have collected the Microsoft Kinect depth images of elderly fall event. After collecting the necessary depth images the background subtraction algorithm has been used to subtract the background and retain the subject. Segmentation and feature selection process have been applied on various daily activity to train the fall detection model. The model has been train using decision tree. To ensure the fall confidence, ground truthing technique has been used.