Improving age measurement in low- and middleincome countries through computer vision: A test in Senegal

Stephane Helleringer, Chong You, Laurence Fleury, Laetitia Douillot, Insa Diouf, Cheikh Tidiane Ndiaye, Valerie Delaunay, Rene Vidal

Research output: Contribution to journalArticle

Abstract

BACKGROUND Age misreporting is pervasive in most low- and middle-income countries (LMIC). It may bias estimates of key demographic indicators, such as those required to track progress towards sustainable development goals. Existing methods to improve age data are often ineffective, cannot be adopted on a large scale, and/or do not permit estimating age over the entire life course. OBJECTIVE We tested a computer vision approach, which produces an age estimate by analyzing a photograph of an individual's face. METHODS We constituted a small training dataset in a population of Senegal covered by a health and demographic surveillance system (HDSS) since 1962. We collected facial images of 353 women aged 18 and above, whose age could be ascertained precisely using HDSS data. We developed automatic age estimation (AAE) systems through machine learning and cross-validation. RESULTS AAE was highly accurate in distinguishing women of reproductive age from women aged 50 and older (area under the curve > 0.95). It allowed estimating age in completed years, with a level of precision comparable to those obtained in European or East Asian populations with training datasets of similar sizes (mean absolute error = 4.62 years). CONCLUSION Computer vision might help improve age ascertainment in demographic datasets collected in LMICs. Further improving the accuracy of this approach will require constituting larger and more complete training datasets in additional LMIC populations. CONTRIBUTION Our work highlights the potential benefits of widely used computer science tools for improving demographic measurement in LMIC settings with deficient data.

Original languageEnglish (US)
Pages (from-to)219-260
Number of pages42
JournalDemographic Research
Volume40
DOIs
StatePublished - Jan 1 2019

Fingerprint

Senegal
income
surveillance
image of women
health
computer science
sustainable development

ASJC Scopus subject areas

  • Demography

Cite this

Improving age measurement in low- and middleincome countries through computer vision : A test in Senegal. / Helleringer, Stephane; You, Chong; Fleury, Laurence; Douillot, Laetitia; Diouf, Insa; Ndiaye, Cheikh Tidiane; Delaunay, Valerie; Vidal, Rene.

In: Demographic Research, Vol. 40, 01.01.2019, p. 219-260.

Research output: Contribution to journalArticle

Helleringer, Stephane ; You, Chong ; Fleury, Laurence ; Douillot, Laetitia ; Diouf, Insa ; Ndiaye, Cheikh Tidiane ; Delaunay, Valerie ; Vidal, Rene. / Improving age measurement in low- and middleincome countries through computer vision : A test in Senegal. In: Demographic Research. 2019 ; Vol. 40. pp. 219-260.
@article{3a6cd4ec96bf4783af068bac0ad72fa0,
title = "Improving age measurement in low- and middleincome countries through computer vision: A test in Senegal",
abstract = "BACKGROUND Age misreporting is pervasive in most low- and middle-income countries (LMIC). It may bias estimates of key demographic indicators, such as those required to track progress towards sustainable development goals. Existing methods to improve age data are often ineffective, cannot be adopted on a large scale, and/or do not permit estimating age over the entire life course. OBJECTIVE We tested a computer vision approach, which produces an age estimate by analyzing a photograph of an individual's face. METHODS We constituted a small training dataset in a population of Senegal covered by a health and demographic surveillance system (HDSS) since 1962. We collected facial images of 353 women aged 18 and above, whose age could be ascertained precisely using HDSS data. We developed automatic age estimation (AAE) systems through machine learning and cross-validation. RESULTS AAE was highly accurate in distinguishing women of reproductive age from women aged 50 and older (area under the curve > 0.95). It allowed estimating age in completed years, with a level of precision comparable to those obtained in European or East Asian populations with training datasets of similar sizes (mean absolute error = 4.62 years). CONCLUSION Computer vision might help improve age ascertainment in demographic datasets collected in LMICs. Further improving the accuracy of this approach will require constituting larger and more complete training datasets in additional LMIC populations. CONTRIBUTION Our work highlights the potential benefits of widely used computer science tools for improving demographic measurement in LMIC settings with deficient data.",
author = "Stephane Helleringer and Chong You and Laurence Fleury and Laetitia Douillot and Insa Diouf and Ndiaye, {Cheikh Tidiane} and Valerie Delaunay and Rene Vidal",
year = "2019",
month = "1",
day = "1",
doi = "10.4054/DEMRES.2019.40.9",
language = "English (US)",
volume = "40",
pages = "219--260",
journal = "Demographic Research",
issn = "1435-9871",
publisher = "Max-Planck Institute for Demographic Research/Max-Planck-institut fur Demografische Forschung",

}

TY - JOUR

T1 - Improving age measurement in low- and middleincome countries through computer vision

T2 - A test in Senegal

AU - Helleringer, Stephane

AU - You, Chong

AU - Fleury, Laurence

AU - Douillot, Laetitia

AU - Diouf, Insa

AU - Ndiaye, Cheikh Tidiane

AU - Delaunay, Valerie

AU - Vidal, Rene

PY - 2019/1/1

Y1 - 2019/1/1

N2 - BACKGROUND Age misreporting is pervasive in most low- and middle-income countries (LMIC). It may bias estimates of key demographic indicators, such as those required to track progress towards sustainable development goals. Existing methods to improve age data are often ineffective, cannot be adopted on a large scale, and/or do not permit estimating age over the entire life course. OBJECTIVE We tested a computer vision approach, which produces an age estimate by analyzing a photograph of an individual's face. METHODS We constituted a small training dataset in a population of Senegal covered by a health and demographic surveillance system (HDSS) since 1962. We collected facial images of 353 women aged 18 and above, whose age could be ascertained precisely using HDSS data. We developed automatic age estimation (AAE) systems through machine learning and cross-validation. RESULTS AAE was highly accurate in distinguishing women of reproductive age from women aged 50 and older (area under the curve > 0.95). It allowed estimating age in completed years, with a level of precision comparable to those obtained in European or East Asian populations with training datasets of similar sizes (mean absolute error = 4.62 years). CONCLUSION Computer vision might help improve age ascertainment in demographic datasets collected in LMICs. Further improving the accuracy of this approach will require constituting larger and more complete training datasets in additional LMIC populations. CONTRIBUTION Our work highlights the potential benefits of widely used computer science tools for improving demographic measurement in LMIC settings with deficient data.

AB - BACKGROUND Age misreporting is pervasive in most low- and middle-income countries (LMIC). It may bias estimates of key demographic indicators, such as those required to track progress towards sustainable development goals. Existing methods to improve age data are often ineffective, cannot be adopted on a large scale, and/or do not permit estimating age over the entire life course. OBJECTIVE We tested a computer vision approach, which produces an age estimate by analyzing a photograph of an individual's face. METHODS We constituted a small training dataset in a population of Senegal covered by a health and demographic surveillance system (HDSS) since 1962. We collected facial images of 353 women aged 18 and above, whose age could be ascertained precisely using HDSS data. We developed automatic age estimation (AAE) systems through machine learning and cross-validation. RESULTS AAE was highly accurate in distinguishing women of reproductive age from women aged 50 and older (area under the curve > 0.95). It allowed estimating age in completed years, with a level of precision comparable to those obtained in European or East Asian populations with training datasets of similar sizes (mean absolute error = 4.62 years). CONCLUSION Computer vision might help improve age ascertainment in demographic datasets collected in LMICs. Further improving the accuracy of this approach will require constituting larger and more complete training datasets in additional LMIC populations. CONTRIBUTION Our work highlights the potential benefits of widely used computer science tools for improving demographic measurement in LMIC settings with deficient data.

UR - http://www.scopus.com/inward/record.url?scp=85065042364&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85065042364&partnerID=8YFLogxK

U2 - 10.4054/DEMRES.2019.40.9

DO - 10.4054/DEMRES.2019.40.9

M3 - Article

AN - SCOPUS:85065042364

VL - 40

SP - 219

EP - 260

JO - Demographic Research

JF - Demographic Research

SN - 1435-9871

ER -