Deep Activity Recognition based on Patterns Discovery for Healthcare Monitoring

Abstract

Healthcare monitoring for humans is important due to several factors including life quality and early detection of health-related problems. Human activity patterns recognition is the most promising ways to monitor human health. Uprisings in the human activity patterns recognition has enabled researchers to recognize multiple health issues through the usage of multiple sensory devices such as motion-based wearable sensors. Irrelevant motion patterns can lead to overlook the important activity recognition in daily living. For this purpose, an early discovery of motion patterns has been proposed for activity recognition in this paper. Main objective is to support the activity detection through motion patterns and deep learning mechanism. The proposed method contains three layered architecture including pre-processing layer, features engineering layer, and classification layer. The anticipated study is investigated over an openly available dataset named Opportunity and results have shown improvement in terms of achieving higher accuracy rate of 88.57%.