@article{fdi:010075176,
  title = {{I}mproving age measurement in low- and middle-income countries through computer vision : a test in {S}enegal},
  author = {{H}elleringer, {S}. and {Y}ou, {C}. and {F}leury, {L}aurence and {D}ouillot, {L}. and {D}iouf, {I}. and {N}diaye, {C}. {T}. and {D}elaunay, {V}al{\'e}rie and {V}idal, {R}.},
  editor = {},
  language = {{ENG}},
  abstract = {{BACKGROUND} {A}ge misreporting is pervasive in most low-and middle-income countries ({LMIC}). {I}t may bias estimates of key demographic indicators, such as those required to track progress towards sustainable development goals. {E}xisting 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} {W}e tested a computer vision approach, which produces an age estimate by analyzing a photograph of an individual's face. {METHODS} {W}e constituted a small training dataset in a population of {S}enegal covered by a health and demographic surveillance system ({HDSS}) since 1962. {W}e collected facial images of 353 women aged 18 and above, whose age could be ascertained precisely using {HDSS} data. {W}e 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). {I}t allowed estimating age in completed years, with a level of precision comparable to those obtained in {E}uropean or {E}ast {A}sian populations with training datasets of similar sizes (mean absolute error = 4.62 years). {CONCLUSION} {C}omputer vision might help improve age ascertainment in demographic datasets collected in {LMIC}s. {F}urther improving the accuracy of this approach will require constituting larger and more complete training datasets in additional {LMIC} populations. {CONTRIBUTION} {O}ur work highlights the potential benefits of widely used computer science tools for improving demographic measurement in {LMIC} settings with deficient data.},
  keywords = {{SENEGAL}},
  booktitle = {},
  journal = {{D}emographic {R}esearch},
  volume = {40},
  numero = {},
  pages = {219--260},
  ISSN = {1435-9871},
  year = {2019},
  DOI = {10.4054/{D}em{R}es.2019.40.9},
  URL = {https://www.documentation.ird.fr/hor/fdi:010075176},
}