Multi-occupancy Fall Detection using Non-Invasive Thermal Vision Sensor

Cankun Zhong; Wing Ng; Shuai Zhang; Chris Nugent; Colin Shewell; Javier Medina-Quero

doi:10.1109/JSEN.2020.3032728

Multi-occupancy Fall Detection using Non-Invasive Thermal Vision Sensor

Cankun Zhong, Wing Ng, Shuai Zhang, Chris Nugent, Colin Shewell, Javier Medina-Quero

Research output: Contribution to journal › Article › peer-review

30 Citations (Scopus)

295 Downloads (Pure)

Abstract

Falling is a common issue within the aging
population. The immediate detection of a fall is key to
guarantee early and immediate attention to avoid other
potential immobility risks and reduction in recovery
time. Video-based approaches for monitoring fall
detection, although being highly accurate, are largely
perceived as being intrusive if deployed within living
environments. As an alternative, thermal vision-based
methods can be deployed to offer a more acceptable
level of privacy. To date, thermal vision-based fall
detection methods have largely focused on
single-occupancy scenarios, which are not fully
representative of real living environments with
multi-occupancy. This work proposes a non-invasive thermal vision-based approach of multi-occupancy fall detection
(MoT-LoGNN) which discriminates between a fall or no-fall. The approach consists of four major components: i) a
multi-occupancy decomposer, ii) a sensitivity-based sample selector, iii) the T-LoGNN for single-occupancy fall
detection, and iv) a fine-tuning mechanism. The T-LoGNN consists of a robust neural network minimizing a Localized
Generalization Error (L-GEM) and thermal image features extracted by a Convolutional Neural Network (CNN).
Comparing to other methods, the MoT-LoGNN achieved the highest average accuracy of 98.39% within the context of a
multi-occupancy fall detection experiment.

Original language	English
Article number	9234482
Pages (from-to)	5377-5388
Number of pages	12
Journal	IEEE Sensors Journal
Volume	21
Issue number	4
Early online date	21 Oct 2020
DOIs	https://doi.org/10.1109/JSEN.2020.3032728
Publication status	Published (in print/issue) - 15 Feb 2021

Bibliographical note

Funding Information:
Manuscript received September 18, 2020; accepted October 14, 2020. Date of publication October 21, 2020; date of current version January 15, 2021. This work was supported in part by the 2020 Research and Development Program in Key Areas of Guangdong Province under Grant 2020B010166002, in part by the National Natural Science Foundation of China under Grant 61876066, in part by the Guangdong Province Science and Technology Plan Project (Collaborative Innovation and Platform Environment Construction) under Grant 2019A050510006, and the REMIND project funded by Marie Sklodowska-Curie EU Framework for Research and Innovation Horizon 2020 under Grant Agreement No. 734355. The associate editor coordinating the review of this article and approving it for publication was Prof. Elena Gaura. (Corresponding author: Wing W. Y. Ng.) Cankun Zhong and Wing W. Y. Ng are with the Guangdong Provincial Key Laboratory of Computational Intelligence and Cyberspace Information, School of Computer Science and Engineering, South China University of Technology, Guangzhou 510006, China (e-mail: [email protected]; [email protected]).

Copyright:
Copyright 2021 Elsevier B.V., All rights reserved.

Keywords

Feature extraction
Image sensors
MoT-LoGNN
Multi-occupancy Fall Detection
Neural Networks
Perturbation methods
Sensors
Temperature sensors
Thermal Vision Sensor
Thermal sensors
Training
smart environments

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1109/JSEN.2020.3032728

09234482Author Accepted Manuscript, 1.75 MB

Cite this

@article{744bb5f2989743fd914583a7bf9f5b61,

title = "Multi-occupancy Fall Detection using Non-Invasive Thermal Vision Sensor",

abstract = "Falling is a common issue within the agingpopulation. The immediate detection of a fall is key toguarantee early and immediate attention to avoid otherpotential immobility risks and reduction in recoverytime. Video-based approaches for monitoring falldetection, although being highly accurate, are largelyperceived as being intrusive if deployed within livingenvironments. As an alternative, thermal vision-basedmethods can be deployed to offer a more acceptablelevel of privacy. To date, thermal vision-based falldetection methods have largely focused onsingle-occupancy scenarios, which are not fullyrepresentative of real living environments withmulti-occupancy. This work proposes a non-invasive thermal vision-based approach of multi-occupancy fall detection(MoT-LoGNN) which discriminates between a fall or no-fall. The approach consists of four major components: i) amulti-occupancy decomposer, ii) a sensitivity-based sample selector, iii) the T-LoGNN for single-occupancy falldetection, and iv) a fine-tuning mechanism. The T-LoGNN consists of a robust neural network minimizing a LocalizedGeneralization Error (L-GEM) and thermal image features extracted by a Convolutional Neural Network (CNN).Comparing to other methods, the MoT-LoGNN achieved the highest average accuracy of 98.39% within the context of amulti-occupancy fall detection experiment.",

keywords = "Feature extraction, Image sensors, MoT-LoGNN, Multi-occupancy Fall Detection, Neural Networks, Perturbation methods, Sensors, Temperature sensors, Thermal Vision Sensor, Thermal sensors, Training, smart environments",

author = "Cankun Zhong and Wing Ng and Shuai Zhang and Chris Nugent and Colin Shewell and Javier Medina-Quero",

note = "Funding Information: Manuscript received September 18, 2020; accepted October 14, 2020. Date of publication October 21, 2020; date of current version January 15, 2021. This work was supported in part by the 2020 Research and Development Program in Key Areas of Guangdong Province under Grant 2020B010166002, in part by the National Natural Science Foundation of China under Grant 61876066, in part by the Guangdong Province Science and Technology Plan Project (Collaborative Innovation and Platform Environment Construction) under Grant 2019A050510006, and the REMIND project funded by Marie Sklodowska-Curie EU Framework for Research and Innovation Horizon 2020 under Grant Agreement No. 734355. The associate editor coordinating the review of this article and approving it for publication was Prof. Elena Gaura. (Corresponding author: Wing W. Y. Ng.) Cankun Zhong and Wing W. Y. Ng are with the Guangdong Provincial Key Laboratory of Computational Intelligence and Cyberspace Information, School of Computer Science and Engineering, South China University of Technology, Guangzhou 510006, China (e-mail: [email protected]; [email protected]). Copyright: Copyright 2021 Elsevier B.V., All rights reserved.",

year = "2021",

month = feb,

day = "15",

doi = "10.1109/JSEN.2020.3032728",

language = "English",

volume = "21",

pages = "5377--5388",

journal = "IEEE Sensors Journal",

issn = "1530-437X",

publisher = "IEEE",

number = "4",

}

TY - JOUR

T1 - Multi-occupancy Fall Detection using Non-Invasive Thermal Vision Sensor

AU - Zhong, Cankun

AU - Ng, Wing

AU - Zhang, Shuai

AU - Nugent, Chris

AU - Shewell, Colin

AU - Medina-Quero, Javier

N1 - Funding Information: Manuscript received September 18, 2020; accepted October 14, 2020. Date of publication October 21, 2020; date of current version January 15, 2021. This work was supported in part by the 2020 Research and Development Program in Key Areas of Guangdong Province under Grant 2020B010166002, in part by the National Natural Science Foundation of China under Grant 61876066, in part by the Guangdong Province Science and Technology Plan Project (Collaborative Innovation and Platform Environment Construction) under Grant 2019A050510006, and the REMIND project funded by Marie Sklodowska-Curie EU Framework for Research and Innovation Horizon 2020 under Grant Agreement No. 734355. The associate editor coordinating the review of this article and approving it for publication was Prof. Elena Gaura. (Corresponding author: Wing W. Y. Ng.) Cankun Zhong and Wing W. Y. Ng are with the Guangdong Provincial Key Laboratory of Computational Intelligence and Cyberspace Information, School of Computer Science and Engineering, South China University of Technology, Guangzhou 510006, China (e-mail: [email protected]; [email protected]). Copyright: Copyright 2021 Elsevier B.V., All rights reserved.

PY - 2021/2/15

Y1 - 2021/2/15

N2 - Falling is a common issue within the agingpopulation. The immediate detection of a fall is key toguarantee early and immediate attention to avoid otherpotential immobility risks and reduction in recoverytime. Video-based approaches for monitoring falldetection, although being highly accurate, are largelyperceived as being intrusive if deployed within livingenvironments. As an alternative, thermal vision-basedmethods can be deployed to offer a more acceptablelevel of privacy. To date, thermal vision-based falldetection methods have largely focused onsingle-occupancy scenarios, which are not fullyrepresentative of real living environments withmulti-occupancy. This work proposes a non-invasive thermal vision-based approach of multi-occupancy fall detection(MoT-LoGNN) which discriminates between a fall or no-fall. The approach consists of four major components: i) amulti-occupancy decomposer, ii) a sensitivity-based sample selector, iii) the T-LoGNN for single-occupancy falldetection, and iv) a fine-tuning mechanism. The T-LoGNN consists of a robust neural network minimizing a LocalizedGeneralization Error (L-GEM) and thermal image features extracted by a Convolutional Neural Network (CNN).Comparing to other methods, the MoT-LoGNN achieved the highest average accuracy of 98.39% within the context of amulti-occupancy fall detection experiment.

AB - Falling is a common issue within the agingpopulation. The immediate detection of a fall is key toguarantee early and immediate attention to avoid otherpotential immobility risks and reduction in recoverytime. Video-based approaches for monitoring falldetection, although being highly accurate, are largelyperceived as being intrusive if deployed within livingenvironments. As an alternative, thermal vision-basedmethods can be deployed to offer a more acceptablelevel of privacy. To date, thermal vision-based falldetection methods have largely focused onsingle-occupancy scenarios, which are not fullyrepresentative of real living environments withmulti-occupancy. This work proposes a non-invasive thermal vision-based approach of multi-occupancy fall detection(MoT-LoGNN) which discriminates between a fall or no-fall. The approach consists of four major components: i) amulti-occupancy decomposer, ii) a sensitivity-based sample selector, iii) the T-LoGNN for single-occupancy falldetection, and iv) a fine-tuning mechanism. The T-LoGNN consists of a robust neural network minimizing a LocalizedGeneralization Error (L-GEM) and thermal image features extracted by a Convolutional Neural Network (CNN).Comparing to other methods, the MoT-LoGNN achieved the highest average accuracy of 98.39% within the context of amulti-occupancy fall detection experiment.

KW - Feature extraction

KW - Image sensors

KW - MoT-LoGNN

KW - Multi-occupancy Fall Detection

KW - Neural Networks

KW - Perturbation methods

KW - Sensors

KW - Temperature sensors

KW - Thermal Vision Sensor

KW - Thermal sensors

KW - Training

KW - smart environments

UR - https://pure.ulster.ac.uk/en/publications/multi-occupancy-fall-detection-using-non-invasive-thermal-vision-

UR - http://www.scopus.com/inward/record.url?scp=85098748073&partnerID=8YFLogxK

U2 - 10.1109/JSEN.2020.3032728

DO - 10.1109/JSEN.2020.3032728

M3 - Article

SN - 1530-437X

VL - 21

SP - 5377

EP - 5388

JO - IEEE Sensors Journal

JF - IEEE Sensors Journal

IS - 4

M1 - 9234482

ER -

Multi-occupancy Fall Detection using Non-Invasive Thermal Vision Sensor

Abstract

Bibliographical note

Keywords

UN SDGs

Access to Document

Other files and links

Fingerprint

Shuai Zhang

Cite this

Multi-occupancy Fall Detection using Non-Invasive Thermal Vision Sensor

Abstract

Bibliographical note

Keywords

UN SDGs

Access to Document

Other files and links

Fingerprint

Profiles

Shuai Zhang

Cite this