Eye Blinking Alarm System for Elderly Abuse Prevention

Abstract

Elderly abuse is an act committed by entrusted person or family member toward elderly people. Elderly abuse is a very serious act and goes against all kind of human rights. Covid-19 era has shown huge increase of abuse committed against elderly people. In the last two years due to the pandemic and quarantined imposed by the government, elderly people had fewer chance to defend themselves against abuse, and less opportunities to alert someone to help, which resulted in increase of elderly abuse cases all over the world. This paper projects the impressive impact of utilizing technology on elderly abuse prevention. This is a potential solution for minimizing the risk of elderly abuse in a cost-effective base. The operation of this project is based on the Eye blinking detection technique with the aid of Eye Aspect Ratio. The detected blinks have been modified to distinguish between voluntary blinks and involuntary blinks and the alarm will be sent by SMS to the elderly’s guardian. The project has been designed using several libraries such as OpenCV, NumPy and Dlib. The project focused on two criteria, distance and brightness. The system was put on trial for 120 times, and as result a detection accuracy of 78.33% has been achieved. The detection accuracy results can be improved by using a camera with higher resolution and fps.