A Robust Multimodal Fall Detection Method for Ambient Assisted Living Applications

Abstract

Accidental falls threaten the lives of people over 65 years of age and can be overcome with quick action for saving lives. Old people who live alone and those who have chronic diseases constitute the main risk groups. Fast and effective detection of falls will increase the quality of life of these people. In this study, using accelerometers together with a video sensor, a multi-modal fall detection mechanism is proposed and its performance has been evaluated. The results indicate that an accelerometer triggered video processing method will minimize the processing costs together with privacy related issues.