Volume 40 - Article 9 | Pages 219–260  

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

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

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).

Conclusions: 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's Affiliation

• Stephane Helleringer - New York University, Abu Dhabi, United Arab Emirates EMAIL
• Chong You - Johns Hopkins University, United States of America EMAIL
• Laurence Fleury - Institut de Recherche pour le Développement (IRD), France EMAIL
• Laetitia Douillot - Institut de Recherche pour le Développement (IRD), France EMAIL
• Insa Diouf - Institut de Recherche pour le Développement (IRD), France EMAIL
• Cheikh Tidiane Ndiaye - Agence Nationale de la Statistique et de la Démographie, Senegal EMAIL
• Valerie Delaunay - Institut de Recherche pour le Développement (IRD), France EMAIL
• Rene Vidal - Johns Hopkins University, United States of America EMAIL

Other articles by the same author/authors in Demographic Research

Estimating mortality from external causes using data from retrospective surveys: A validation study in Niakhar (Senegal)
Volume 38 - Article 32

Excess mortality associated with HIV: Survey estimates from the PHIA project
Volume 51 - Article 38

Knowledge, risk perceptions, and behaviors related to the COVID-19 pandemic in Malawi
Volume 44 - Article 20

The Likoma Network Study: Context, data collection and initial results
Volume 21 - Article 15

Most recent similar articles in Demographic Research

Is single parenthood increasingly an experience of less-educated mothers? A European comparison over five decades
Volume 51 - Article 34    | Keywords: age, children, cross-national comparison, education, Europe, family life course, inequality, single motherhood

Data errors in mortality estimation: Formal demographic analysis of under-registration, under-enumeration, and age misreporting
Volume 51 - Article 9    | Keywords: age misreporting, data errors, formal demography, mortality

A Bayesian model for the reconstruction of education- and age-specific fertility rates: An application to African and Latin American countries
Volume 49 - Article 31    | Keywords: age, Bayesian analysis, education, fertility estimation, fertility rates

Age reporting for the oldest old in the Brazilian COVID-19 vaccination database: What can we learn from it?
Volume 48 - Article 28    | Keywords: age misreporting, Brazil, COVID-19, mortality crossover, oldest old, population aging, vaccinations

Better to ask online when it concerns intimate relationships? Survey mode differences in the assessment of relationship quality
Volume 48 - Article 22    | Keywords: data collection, Generations and Gender Programme (GGP), interviews, intimate relationship, quality of a statistical survey, relationships, research methods, survey data