Detection of Fall Event with Extreme Learning Machine and One Dimensional Local Binary Pattern

Abstract

We can use developing technology to make people's lives easier and to respond to their needs accurately and effectively. Therefore, the focus of most of the work is on people and facilitating people's lives. Many studies focusing on people have examined human movements; human movements were detected and identified. The act of falling between different human movements can cause serious vital problems for people, especially for the elderly. When the fall occurs, first aid must be delivered to the falling individual quickly. We can use technology to accurately and effectively detect falls and provide emergency first aid to falling individuals. In this study, we examined human movements for effective and efficient detection of fall action, and we performed fall detection using one-dimensional LBP (1D-Local Binary Pattern) and ELM (Extreme Leaning Machine) method. Although only acceleration data is used in our application, these methods have achieved very good results and these results are presented at the end of the article. © 2019 IEEE.