Real-Time Posture Estimation-Based Human Fall Detection System

Abstract

Special attention should be given to those in vulnerable age groups, such as children and the elderly, who live in solitary living arrangements and have an increased likelihood of experiencing falls. The primary objective of a fall detector is to minimize the duration of time an elderly person remains on the ground subsequent to experiencing a fall event. The length of time that an individual remains on the floor subsequent to a fall is a determining factor in assessing the severity of the incident. The act of engaging in prolonged deceit has been shown to elevate the likelihood of experiencing adverse health outcomes such as hypothermia, dehydration, and pressure ulcers. The ultimate objective of the detector system is to identify the occurrence of a fall and expeditiously notify a carer. The potential for individuals to inadvertently trigger a fall detector by engaging in rapid movements such as standing up and sitting down underscores the need for a dependable detector that can effectively differentiate between instances of falling and other related occurrences. The primary aim of this research is to enhance the understanding of fall-related mechanisms among medical technologists in the field of public health. The assessment of an individual's physique alteration potential relies on the calculation of the ratio between their height and width. To validate a human fall, it is necessary to measure the height and centre of the rectangle that encloses the individual, and thereafter compare these measurements to a specified threshold. An alert system has been developed to notify persons who are connected to the network in the event of a catastrophe, provided that a state of inactivity is maintained for 100 consecutive frames.