Human fall detection for video surveillance by handling partial occlusion scenario

Video Surveillance is a usual topic when it comes to enhancing security and safety in the intelligent home environments. With the advancement of technology in medical science over past decades, and large amount of increment in the population of elderly people. Falls is the one of main cause of injuries among the elderly people. Therefore, there is urgent need of such surveillance systems that are able to send a cell/message/alarm for help, in the case of some incident happens where a person slips and falls and is unable to call for help i.e. he/she loses consciousness. The proposed fall detection system is based on a change in human shape in daily activities. In the proposed approach, features are extracted from the human silhouette and a fall is detected by analyzing the change in its shape. The proposed approach also handles the problem of partial occlusion during the fall detection. The approach shows satisfactory results as compared to the sate-of-art method.