Proposal and verification of a foot arch classification technique using foot pressure distribution images

Yumi Iwakami, Emi Anzai, Kanako Nakajima, Kazuya Imaizumi, Kazuhiko Yamashita, Makoto Okabe, Rikio Onai

Research output: Contribution to journalArticle

Abstract

The shape of our feet is important for our good health. If the feet are an abnormal shape, such as 'high arch' and 'flat foot', it may adversely affect the performance of walking and maintaining the posture, and damage our health. Therefore, we want to develop a system that allows even non-expert users to measure the shape of their feet and make a diagnosis. Our system takes the image of pressure distribution of feet as an input, and classifies it into four categories: 'normal', 'high arch', 'flat foot', and 'suspected of an abnormal shape'. We build our classifier using the Adaboost algorithm and decision tree algorithm. Our training data consists of 200 images of pressure distribution of feet that are labeled by the experts in advance. We use the pressure and area information obtained from each input image as the image features. We conduct the verification based on the cross validation on our training data. The resulting accuracy that our classifier achieves is 95.5%. The recalls of the four categories of 'high arch', 'normal', 'flat foot', and 'suspected of an abnormal shape' are 100%, 94.2%, 96.3%, and 95.8%, which are best performance compared with the previous methods.

Original languageEnglish
Pages (from-to)505-512
Number of pages8
JournalIEEJ Transactions on Electronics, Information and Systems
Volume135
Issue number5
DOIs
Publication statusPublished - 2015 May 1
Externally publishedYes

Fingerprint

Arches
Pressure distribution
Classifiers
Health
Adaptive boosting
Decision trees

Keywords

  • Foot arch classification
  • Foot pressure distribution images
  • Machine learning

ASJC Scopus subject areas

  • Electrical and Electronic Engineering

Cite this

Proposal and verification of a foot arch classification technique using foot pressure distribution images. / Iwakami, Yumi; Anzai, Emi; Nakajima, Kanako; Imaizumi, Kazuya; Yamashita, Kazuhiko; Okabe, Makoto; Onai, Rikio.

In: IEEJ Transactions on Electronics, Information and Systems, Vol. 135, No. 5, 01.05.2015, p. 505-512.

Research output: Contribution to journalArticle

Iwakami, Yumi ; Anzai, Emi ; Nakajima, Kanako ; Imaizumi, Kazuya ; Yamashita, Kazuhiko ; Okabe, Makoto ; Onai, Rikio. / Proposal and verification of a foot arch classification technique using foot pressure distribution images. In: IEEJ Transactions on Electronics, Information and Systems. 2015 ; Vol. 135, No. 5. pp. 505-512.
@article{3337433812794f69b2b9451386fb3911,
title = "Proposal and verification of a foot arch classification technique using foot pressure distribution images",
abstract = "The shape of our feet is important for our good health. If the feet are an abnormal shape, such as 'high arch' and 'flat foot', it may adversely affect the performance of walking and maintaining the posture, and damage our health. Therefore, we want to develop a system that allows even non-expert users to measure the shape of their feet and make a diagnosis. Our system takes the image of pressure distribution of feet as an input, and classifies it into four categories: 'normal', 'high arch', 'flat foot', and 'suspected of an abnormal shape'. We build our classifier using the Adaboost algorithm and decision tree algorithm. Our training data consists of 200 images of pressure distribution of feet that are labeled by the experts in advance. We use the pressure and area information obtained from each input image as the image features. We conduct the verification based on the cross validation on our training data. The resulting accuracy that our classifier achieves is 95.5{\%}. The recalls of the four categories of 'high arch', 'normal', 'flat foot', and 'suspected of an abnormal shape' are 100{\%}, 94.2{\%}, 96.3{\%}, and 95.8{\%}, which are best performance compared with the previous methods.",
keywords = "Foot arch classification, Foot pressure distribution images, Machine learning",
author = "Yumi Iwakami and Emi Anzai and Kanako Nakajima and Kazuya Imaizumi and Kazuhiko Yamashita and Makoto Okabe and Rikio Onai",
year = "2015",
month = "5",
day = "1",
doi = "10.1541/ieejeiss.135.505",
language = "English",
volume = "135",
pages = "505--512",
journal = "IEEJ Transactions on Electronics, Information and Systems",
issn = "0385-4221",
publisher = "The Institute of Electrical Engineers of Japan",
number = "5",

}

TY - JOUR

T1 - Proposal and verification of a foot arch classification technique using foot pressure distribution images

AU - Iwakami, Yumi

AU - Anzai, Emi

AU - Nakajima, Kanako

AU - Imaizumi, Kazuya

AU - Yamashita, Kazuhiko

AU - Okabe, Makoto

AU - Onai, Rikio

PY - 2015/5/1

Y1 - 2015/5/1

N2 - The shape of our feet is important for our good health. If the feet are an abnormal shape, such as 'high arch' and 'flat foot', it may adversely affect the performance of walking and maintaining the posture, and damage our health. Therefore, we want to develop a system that allows even non-expert users to measure the shape of their feet and make a diagnosis. Our system takes the image of pressure distribution of feet as an input, and classifies it into four categories: 'normal', 'high arch', 'flat foot', and 'suspected of an abnormal shape'. We build our classifier using the Adaboost algorithm and decision tree algorithm. Our training data consists of 200 images of pressure distribution of feet that are labeled by the experts in advance. We use the pressure and area information obtained from each input image as the image features. We conduct the verification based on the cross validation on our training data. The resulting accuracy that our classifier achieves is 95.5%. The recalls of the four categories of 'high arch', 'normal', 'flat foot', and 'suspected of an abnormal shape' are 100%, 94.2%, 96.3%, and 95.8%, which are best performance compared with the previous methods.

AB - The shape of our feet is important for our good health. If the feet are an abnormal shape, such as 'high arch' and 'flat foot', it may adversely affect the performance of walking and maintaining the posture, and damage our health. Therefore, we want to develop a system that allows even non-expert users to measure the shape of their feet and make a diagnosis. Our system takes the image of pressure distribution of feet as an input, and classifies it into four categories: 'normal', 'high arch', 'flat foot', and 'suspected of an abnormal shape'. We build our classifier using the Adaboost algorithm and decision tree algorithm. Our training data consists of 200 images of pressure distribution of feet that are labeled by the experts in advance. We use the pressure and area information obtained from each input image as the image features. We conduct the verification based on the cross validation on our training data. The resulting accuracy that our classifier achieves is 95.5%. The recalls of the four categories of 'high arch', 'normal', 'flat foot', and 'suspected of an abnormal shape' are 100%, 94.2%, 96.3%, and 95.8%, which are best performance compared with the previous methods.

KW - Foot arch classification

KW - Foot pressure distribution images

KW - Machine learning

UR - http://www.scopus.com/inward/record.url?scp=84929403296&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84929403296&partnerID=8YFLogxK

U2 - 10.1541/ieejeiss.135.505

DO - 10.1541/ieejeiss.135.505

M3 - Article

VL - 135

SP - 505

EP - 512

JO - IEEJ Transactions on Electronics, Information and Systems

JF - IEEJ Transactions on Electronics, Information and Systems

SN - 0385-4221

IS - 5

ER -