Camera-based toddler fall detection system by using kalman filter

Research output: Contribution to journalArticle

Abstract

Monitoring a toddler is a tedious job, yet a very important one. Fall down is the most common risk that leads to injury during the process of learning to walk. Thus, this paper proposed an algorithm to detect automatically the event of toddler fall down to assist the supervision process by alerting the caretaker if necessary. This system comprises of a background subtraction module to detect region of interest, a tracking module using Kalman filter to track toddler movement and a decision module through decision tree process to determine the toddler state. System performance is evaluated based on three metrics, which are accuracy, sensitivity and specificity. The proposed algorithm works well with a low error performance. Further research should be done to improve the robustness of the system for real life environment implementation.

Original languageEnglish
Pages (from-to)383-388
Number of pages6
JournalJournal of Theoretical and Applied Information Technology
Volume81
Issue number2
Publication statusPublished - 1 Nov 2015

Fingerprint

Kalman filters
Kalman Filter
Camera
Cameras
Module
Decision trees
Background Subtraction
Region of Interest
Decision tree
Walk
Specificity
Monitoring
System Performance
Robustness
Metric
Necessary
Life
Learning
Movement

Keywords

  • Background subtraction module
  • Decision tree process
  • Fall down
  • Kalman filter
  • Toddler monitoring

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

@article{70e8e91c46334c64b43b187518ce5bc6,
title = "Camera-based toddler fall detection system by using kalman filter",
abstract = "Monitoring a toddler is a tedious job, yet a very important one. Fall down is the most common risk that leads to injury during the process of learning to walk. Thus, this paper proposed an algorithm to detect automatically the event of toddler fall down to assist the supervision process by alerting the caretaker if necessary. This system comprises of a background subtraction module to detect region of interest, a tracking module using Kalman filter to track toddler movement and a decision module through decision tree process to determine the toddler state. System performance is evaluated based on three metrics, which are accuracy, sensitivity and specificity. The proposed algorithm works well with a low error performance. Further research should be done to improve the robustness of the system for real life environment implementation.",
keywords = "Background subtraction module, Decision tree process, Fall down, Kalman filter, Toddler monitoring",
author = "Leong, {Low Han} and Aini Hussain and Zulkifley, {Mohd Asyraf} and {Wan Zaki}, {Wan Mimi Diyana}",
year = "2015",
month = "11",
day = "1",
language = "English",
volume = "81",
pages = "383--388",
journal = "Journal of Theoretical and Applied Information Technology",
issn = "1992-8645",
publisher = "Asian Research Publishing Network (ARPN)",
number = "2",

}

TY - JOUR

T1 - Camera-based toddler fall detection system by using kalman filter

AU - Leong, Low Han

AU - Hussain, Aini

AU - Zulkifley, Mohd Asyraf

AU - Wan Zaki, Wan Mimi Diyana

PY - 2015/11/1

Y1 - 2015/11/1

N2 - Monitoring a toddler is a tedious job, yet a very important one. Fall down is the most common risk that leads to injury during the process of learning to walk. Thus, this paper proposed an algorithm to detect automatically the event of toddler fall down to assist the supervision process by alerting the caretaker if necessary. This system comprises of a background subtraction module to detect region of interest, a tracking module using Kalman filter to track toddler movement and a decision module through decision tree process to determine the toddler state. System performance is evaluated based on three metrics, which are accuracy, sensitivity and specificity. The proposed algorithm works well with a low error performance. Further research should be done to improve the robustness of the system for real life environment implementation.

AB - Monitoring a toddler is a tedious job, yet a very important one. Fall down is the most common risk that leads to injury during the process of learning to walk. Thus, this paper proposed an algorithm to detect automatically the event of toddler fall down to assist the supervision process by alerting the caretaker if necessary. This system comprises of a background subtraction module to detect region of interest, a tracking module using Kalman filter to track toddler movement and a decision module through decision tree process to determine the toddler state. System performance is evaluated based on three metrics, which are accuracy, sensitivity and specificity. The proposed algorithm works well with a low error performance. Further research should be done to improve the robustness of the system for real life environment implementation.

KW - Background subtraction module

KW - Decision tree process

KW - Fall down

KW - Kalman filter

KW - Toddler monitoring

UR - http://www.scopus.com/inward/record.url?scp=84947610795&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84947610795&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:84947610795

VL - 81

SP - 383

EP - 388

JO - Journal of Theoretical and Applied Information Technology

JF - Journal of Theoretical and Applied Information Technology

SN - 1992-8645

IS - 2

ER -