An ambient assisted living system for Alzheimer’s patients

Abstract

Alzheimer’s is a brain disorder that gradually deteriorates the brain functions of the patients. As the disease progresses, victims start to lose their memory, thinking ability, eventually rendering them unable to perform basic tasks. They also face many difficulties namely disorientation, wandering, aggression, insomnia, hallucination, etc. What makes the situation worse is that when the caregivers try to help them most of the time they tend not to cooperate. In this paper, we have designed an AI that assists the sufferers in combating these issues by analyzing their environment, daily routine, interests, behavioral patterns, and many more factors. Using computer vision we have created a face recognition framework that identifies individuals in front of the patient & shows him/her their name, how they are related, and some photos & videos of them together. We also used an object detection system that helps prevent wandering by constantly monitoring the surroundings of the patient & notifying the caretakers about items such as keys, shoes, handbags, doors etc that could influence the patient to leave the house. The AI is instructed to alarm the attendant continuously if the patient somehow succeeds to go beyond the safe area. This feature allows the caregivers some free time as they don’t need to monitor the patients 24/7 anymore. The face recognition framework achieves accuracy of 97.44% and the object detection system has mAP of 72.3% that uses YOLOv7 model. Thus, this study tries to achieve its goal to make life comparatively easier for the patients & the caregivers by making the patients self-dependent & discharging the attendants from some of their tasks.