Detection and retrieval of human falls on furniture using hybrid method

Abstract

Automatic detection of human falls is very important to detect accidental falls at Medical places, and at places where elderly and vulnerable people living alone at home. Many methods have been proposed to detect human falls using various approaches, but cannot give accurate results in complex environments. This paper proposes a new hybrid method which detects falls and then retrieve relevant images. The proposed algorithm first calculates the features of hybrid methods and retrieve the similar images from the database using various methods. the proposed method achieves up to 97.5% precision and recall of 73%.