A Mobile Application for Physical Activity Recognition using Acceleration Data from Wearable Sensors for Cardiac Rehabilitation

Abstract

mHealth applications are an ever-expanding frontier in today's use of technology. They allow a user to record health data and contact their doctor from the convenience of a smartphone. This paper presents a first version release of a mobile application that aims to assess compliance of cardiovascular diseased patients with home-based cardiac rehabilitation, by monitoring physical activities using wearable sensors. The application generates reports for both the patient and the doctor through an interactive dashboard, as initial proposal, that provides feedback of physical activities of daily living undertaken by the patient. The application integrates a human activity recognition system, which learns a support vector machine algorithm to identify 10 different daily activities, such as walking, going upstairs, sitting and lying, from accelerometer data using a connected textile including movement sensors. Our early deployment and execution results are promising since they are showing good accuracy for recognizing all the ten daily living activities.

Mobilizing artificial intelligence to cardiac telerehabilitation

“…Identifying various activities of the human body in real time needs an efficient and wellworking model which may involve the application of some AI algorithms. Previous studies have shown the feasibility of using machine learning to design and train an accurate classifier for data feature selection, which enables wearables to capture and recognize various kinds of human activities during cardiac telerehabilitation [59,60]. Fig.…”

“…The reliability of the support vector machine (SVM) classi-fier, one kind of ML classifier, was tested according to the process of Leave-one-subject-out cross validation, resulting in an accuracy rate of 95.4% in classification [59]. In this study, individual features of patients ranked in the top ten were selected to train the SVM model [59]. Features selection is the most crucial aspect of building a highly accurate classifier.…”

“…Table 1 (Ref. [59][60][61]) lists the different features chosen to train models and the feature selection method used in these studies. Extending beyond fitness detection and recognition, an AI server was used to identify and manage abnormal vital signs when patients were doing physical activities of CR at home [62].…”

“…Features: top ten features in time-domain and frequency domain[59]. Feature selection method: Relief-F algorithms[59].CSVMPhysical activity classification.…”

“…Features: top ten features in time-domain and frequency domain[59]. Feature selection method: Relief-F algorithms[59].CSVMPhysical activity classification. Features: seven features were selected finally: co-relation X and Y axis, corelation Y and Z axis, minimum value along X axis, kurtosis of data around X axis, skewness around X axis, standard deviation X axis, the sum of aggregate acceleration Ai (if Ai ≤25th percentile)[60].Feature selection method: 12-fold cross validation.…”