Automated classification of demographics from face images

A tutorial and validation

Research output: Working paperOther research output

28 Downloads (Pure)

Abstract

Examining disparities in social outcomes as a function of gender, age, or race has a long tradition in psychology and related disciplines. With an increasing availability of large naturalistic data sets, researchers are afforded the opportunity to study the effects of demographic characteristics with real-world data and high statistical power. However, since demographic characteristics are often determined by having participants rate images of targets, limits in participant pools can hinder researchers from analyzing large data sets. Here, we present a tutorial on how to use two face classification algorithms, Face++ and Kairos. We also test and compare their accuracy under varying conditions and provide practical recommendations for their use. Drawing on three face databases (n = 3,141 images), we find that classification accuracy of the algorithms is (a) generally high and similar to the accuracy of human raters, (b) similar for standardized and more variable images, and (c) dependent on various factors such as the target’s race, the angle from which targets were photographed, and which algorithm is used. In sum, we propose that automated face classification can be a useful tool for researchers interested in studying the effects of demographic characteristics in large naturalistic data sets.
Original languageEnglish
Publication statusSubmitted - 2019

Fingerprint

Availability

Cite this

@techreport{5506d18a8faa49928cb76cb85fbc1cf1,
title = "Automated classification of demographics from face images: A tutorial and validation",
abstract = "Examining disparities in social outcomes as a function of gender, age, or race has a long tradition in psychology and related disciplines. With an increasing availability of large naturalistic data sets, researchers are afforded the opportunity to study the effects of demographic characteristics with real-world data and high statistical power. However, since demographic characteristics are often determined by having participants rate images of targets, limits in participant pools can hinder researchers from analyzing large data sets. Here, we present a tutorial on how to use two face classification algorithms, Face++ and Kairos. We also test and compare their accuracy under varying conditions and provide practical recommendations for their use. Drawing on three face databases (n = 3,141 images), we find that classification accuracy of the algorithms is (a) generally high and similar to the accuracy of human raters, (b) similar for standardized and more variable images, and (c) dependent on various factors such as the target’s race, the angle from which targets were photographed, and which algorithm is used. In sum, we propose that automated face classification can be a useful tool for researchers interested in studying the effects of demographic characteristics in large naturalistic data sets.",
author = "Bastian Jaeger and Willem Sleegers and Anthony Evans",
year = "2019",
language = "English",
type = "WorkingPaper",

}

TY - UNPB

T1 - Automated classification of demographics from face images

T2 - A tutorial and validation

AU - Jaeger, Bastian

AU - Sleegers, Willem

AU - Evans, Anthony

PY - 2019

Y1 - 2019

N2 - Examining disparities in social outcomes as a function of gender, age, or race has a long tradition in psychology and related disciplines. With an increasing availability of large naturalistic data sets, researchers are afforded the opportunity to study the effects of demographic characteristics with real-world data and high statistical power. However, since demographic characteristics are often determined by having participants rate images of targets, limits in participant pools can hinder researchers from analyzing large data sets. Here, we present a tutorial on how to use two face classification algorithms, Face++ and Kairos. We also test and compare their accuracy under varying conditions and provide practical recommendations for their use. Drawing on three face databases (n = 3,141 images), we find that classification accuracy of the algorithms is (a) generally high and similar to the accuracy of human raters, (b) similar for standardized and more variable images, and (c) dependent on various factors such as the target’s race, the angle from which targets were photographed, and which algorithm is used. In sum, we propose that automated face classification can be a useful tool for researchers interested in studying the effects of demographic characteristics in large naturalistic data sets.

AB - Examining disparities in social outcomes as a function of gender, age, or race has a long tradition in psychology and related disciplines. With an increasing availability of large naturalistic data sets, researchers are afforded the opportunity to study the effects of demographic characteristics with real-world data and high statistical power. However, since demographic characteristics are often determined by having participants rate images of targets, limits in participant pools can hinder researchers from analyzing large data sets. Here, we present a tutorial on how to use two face classification algorithms, Face++ and Kairos. We also test and compare their accuracy under varying conditions and provide practical recommendations for their use. Drawing on three face databases (n = 3,141 images), we find that classification accuracy of the algorithms is (a) generally high and similar to the accuracy of human raters, (b) similar for standardized and more variable images, and (c) dependent on various factors such as the target’s race, the angle from which targets were photographed, and which algorithm is used. In sum, we propose that automated face classification can be a useful tool for researchers interested in studying the effects of demographic characteristics in large naturalistic data sets.

M3 - Working paper

BT - Automated classification of demographics from face images

ER -