Home // eTELEMED 2020, The Twelfth International Conference on eHealth, Telemedicine, and Social Medicine // View article

Detection and Classification Method for a Temporary Change in Walking

Authors:
Shin Morishima
Misato Haruta
Akira Urashima
Tomoji Toriyama

Keywords: Walking recognition; Classification; Anomaly detection; Human activity recognition.

Abstract:
The global elderly population has grown in recent years and one of the difficulties the elderly face is an increased vulnerability to falls. One of the ways to alleviate this problem is to identify actions that cause falls and to prevent falls using the detection result. Temporary change in walking (i.e., stumble and a stagger) is a typical action. However, existing studies only classify walking or other activities and recognize walking speed, and these studies do not focus on a temporary change in walking. In this paper, we propose a detection method for a change in walking using a change point detection (i.e., anomaly detection for the time series data) and a classification method for the multiple types of change. During the evaluation, four types of anomaly walking videos (i.e. anomaly represents a temporary change.) are used (the total number of videos is 240). As a result, our method can detect anomaly walking in 91.7% and classify four types of detected anomaly walking into three clusters in 89.1% on the basis of each characteristic.

Pages: 74 to 79

Copyright: Copyright (c) IARIA, 2020

Publication date: March 22, 2020

Published in: conference

ISSN: 2308-4359

ISBN: 978-1-61208-763-4

Location: Valencia, Spain

Dates: from November 21, 2020 to November 25, 2020