Hierarchical Classification on Multimodal Sensing for Human Activity Recogintion and Fall Detection

Haobo Li; Aman Shrestha; Francesco Fioranelli; Julien Le Kernec; Hadi Heidari

doi:10.1109/ICSENS.2018.8589797

Hierarchical Classification on Multimodal Sensing for Human Activity Recogintion and Fall Detection

Haobo Li
, Aman Shrestha
, Francesco Fioranelli (Lead / Corresponding author)
, Julien Le Kernec
, Hadi Heidari

Biomedical Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

9 Citations (Scopus)

Abstract

This paper presents initial results on the usage of hierarchical classification for human activities discrimination and fall detection in the context of assisted living. Multimodal sensing is proposed by combining data from a wearable device and a radar system. The effect of different approaches in selecting the activities in each sub-group of the hierarchy are explored and reported as preliminary results in this work, while a more detailed investigation is undergoing. 1.2-2.2% improvement in accuracy with SVM and DL classifiers compared with the conventional case of activity classification is reported; subsequent improvement (1.6%) occurs when using SVM-SFS in the second stage of hierarchical classification.

Original language	English
Title of host publication	2018 Ieee Sensors proceedings
Publisher	IEEE
Pages	1059-1062
Number of pages	4
ISBN (Print)	978-1-5386-4707-3
DOIs	https://doi.org/10.1109/ICSENS.2018.8589797
Publication status	Published - 26 Dec 2018
Event	17th IEEE Sensors Conference 2018 - Pullman Aeorcity, New Delhi, India Duration: 28 Oct 2018 → 31 Oct 2018

Conference

Conference	17th IEEE Sensors Conference 2018
Abbreviated title	SENSORS 2018
Country/Territory	India
City	New Delhi
Period	28/10/18 → 31/10/18

Keywords

Fall detection
Hierarchical classification
Human activity recognition
Machine learning
Multi-modal sensing

Access to Document

10.1109/ICSENS.2018.8589797

Cite this

@inproceedings{3ed1c443fe6448088c605af84c398954,

title = "Hierarchical Classification on Multimodal Sensing for Human Activity Recogintion and Fall Detection",

abstract = "This paper presents initial results on the usage of hierarchical classification for human activities discrimination and fall detection in the context of assisted living. Multimodal sensing is proposed by combining data from a wearable device and a radar system. The effect of different approaches in selecting the activities in each sub-group of the hierarchy are explored and reported as preliminary results in this work, while a more detailed investigation is undergoing. 1.2-2.2\% improvement in accuracy with SVM and DL classifiers compared with the conventional case of activity classification is reported; subsequent improvement (1.6\%) occurs when using SVM-SFS in the second stage of hierarchical classification.",

keywords = "Fall detection, Hierarchical classification, Human activity recognition, Machine learning, Multi-modal sensing",

author = "Haobo Li and Aman Shrestha and Francesco Fioranelli and \{Le Kernec\}, Julien and Hadi Heidari",

year = "2018",

month = dec,

day = "26",

doi = "10.1109/ICSENS.2018.8589797",

language = "English",

isbn = "978-1-5386-4707-3",

pages = "1059--1062",

booktitle = "2018 Ieee Sensors proceedings",

publisher = "IEEE",

note = "17th IEEE Sensors Conference 2018, SENSORS 2018 ; Conference date: 28-10-2018 Through 31-10-2018",

}

TY - GEN

T1 - Hierarchical Classification on Multimodal Sensing for Human Activity Recogintion and Fall Detection

AU - Li, Haobo

AU - Shrestha, Aman

AU - Fioranelli, Francesco

AU - Le Kernec, Julien

AU - Heidari, Hadi

PY - 2018/12/26

Y1 - 2018/12/26

N2 - This paper presents initial results on the usage of hierarchical classification for human activities discrimination and fall detection in the context of assisted living. Multimodal sensing is proposed by combining data from a wearable device and a radar system. The effect of different approaches in selecting the activities in each sub-group of the hierarchy are explored and reported as preliminary results in this work, while a more detailed investigation is undergoing. 1.2-2.2% improvement in accuracy with SVM and DL classifiers compared with the conventional case of activity classification is reported; subsequent improvement (1.6%) occurs when using SVM-SFS in the second stage of hierarchical classification.

AB - This paper presents initial results on the usage of hierarchical classification for human activities discrimination and fall detection in the context of assisted living. Multimodal sensing is proposed by combining data from a wearable device and a radar system. The effect of different approaches in selecting the activities in each sub-group of the hierarchy are explored and reported as preliminary results in this work, while a more detailed investigation is undergoing. 1.2-2.2% improvement in accuracy with SVM and DL classifiers compared with the conventional case of activity classification is reported; subsequent improvement (1.6%) occurs when using SVM-SFS in the second stage of hierarchical classification.

KW - Fall detection

KW - Hierarchical classification

KW - Human activity recognition

KW - Machine learning

KW - Multi-modal sensing

UR - https://www.scopus.com/pages/publications/85060846226

UR - https://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=pure_staging_and_production&SrcAuth=WosAPI&KeyUT=WOS:000468199300275&DestLinkType=FullRecord&DestApp=WOS_CPL

U2 - 10.1109/ICSENS.2018.8589797

DO - 10.1109/ICSENS.2018.8589797

M3 - Conference contribution

SN - 978-1-5386-4707-3

SP - 1059

EP - 1062

BT - 2018 Ieee Sensors proceedings

PB - IEEE

T2 - 17th IEEE Sensors Conference 2018

Y2 - 28 October 2018 through 31 October 2018

ER -

Hierarchical Classification on Multimodal Sensing for Human Activity Recogintion and Fall Detection

Abstract

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this