ePrints

Wearable Social Sensing: Content-Based Processing Methodology and Implementation

Lookup NU author(s): Bin Gao, Dr Wai Lok Woo

Downloads

Full text for this publication is not currently held within this repository. Alternative links are provided below where available.

Abstract

IEEE Developing wearable activity and speech sensing for assessing human physical and mental health is just as significant as conscious content for determining social behavior. Multiple social relevant sensors such as microphones and accelerometer embedded in wearable devices paves the way to provide the opportunity to continuously and non-invasively monitor anxiety and stress in real-life situation. In this paper, we present the design, implementation, and deployment of a wearable computing platform capable of automatically extracting and analyzing social signals. In particular, we benchmarked a set of integrated social features to objectively quantify the level of anxiety using an in-house built wearable device. In addition, in order to protect privacy, we propose a potential method to embed the audio features processing in the hardware to avoid recording their voice directly. In addition, we have implemented the k-means classification to determine the level of anxiety of the subjects. The obtained performance has demonstrated that both activity and speech social features have the potential to directly infer anxiety across multiple individuals.

Publication metadata

Author(s): Gu J, Gao B, Chen Y, Jiang L, Gao Z, Ma X, Ma Y, Woo WL, Jin J

Publication type: Article

Publication status: Published

Journal: IEEE Sensors Journal

Year: 2017

Volume: 17

Issue: 21

Pages: 7167-7176

Print publication date: 01/11/2017

Online publication date: 19/09/2017

Acceptance date: 11/09/2017

ISSN (print): 1530-437X

ISSN (electronic): 1558-1748

Publisher: IEEE

URL: https://doi.org/10.1109/JSEN.2017.2754289

DOI: 10.1109/JSEN.2017.2754289

Altmetrics

Share