The Impact of Age and Threshold Variation on Facial Recognition Algorithm Performance Using Images of Children

Michalski, Dana; Yiu, Sau Yee; Malec, Chris

doi:10.1109/icb2018.2018.00041

Cited by 23 publications

(20 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This, obviously, is an aggregate result; it will generally be possible to find some individuals from different age groups who produce high imposter scores but this will be increasingly difficult as the age difference increases. These results are similar to those reported by Michalski et al [28] for false positives in children using one commercial algorithm. The report also shows false negative ageing effects broken out by age at enrolment, and time lapse.…”

Section: Dependence Of Fmr On Race In United States Mugshotssupporting

confidence: 91%

“…A poor image can undermine detection or recognition, and it is possible that certain demographics yield photographs ill-suited to face recognition e.g. young children [28], or very tall individuals. As pointed out above there is potential for demographic differentials to appear at the capture stage, that is when only a single image is being collected before any comparison with other images.…”

Section: The Role Of Image Qualitymentioning

confidence: 99%

See 1 more Smart Citation

Face recognition vendor test part 3:

Grother¹,

Ngan²,

Hanaoka³

2019

238

113

View full text Add to dashboard Cite

This is the third in a series of reports on ongoing face recognition vendor tests (FRVT) executed by the National Institute of Standards and Technology (NIST). The first two reports cover, respectively, the performance of one-to-one face recognition algorithms used for verification of asserted identities, and performance of one-to-many face recognition algorithms used for identification of individuals in photo data bases. This document extends those evaluations to document accuracy variations across demographic groups. MOTIVATION The recent expansion in the availability, capability, and use of face recognition has been accompanied by assertions that demographic dependencies could lead to accuracy variations and potential bias. A report from Georgetown University [14] work noted that prior studies [22], articulated sources of bias, described the potential impacts particularly in a policing context, and discussed policy and regulatory implications. Additionally, this work is motivated by studies of demographic effects in more recent face recognition [9, 16, 23] and gender estimation algorithms [5, 36]. AIMS AND SCOPE NIST has conducted tests to quantify demographic differences in contemporary face recognition algorithms. This report provides details about the recognition process, notes where demographic effects could occur, details specific performance metrics and analyses, gives empirical results, and recommends research into the mitigation of performance deficiencies. NIST intends this report to inform discussion and decisions about the accuracy, utility, and limitations of face recognition technologies. Its intended audience includes policy makers, face recognition algorithm developers, systems integrators, and managers of face recognition systems concerned with mitigation of risks implied by demographic differentials.

show abstract

Section: Dependence Of Fmr On Race In United States Mugshotssupporting

confidence: 91%

Section: The Role Of Image Qualitymentioning

confidence: 99%

Face recognition vendor test part 3:

Grother¹,

Ngan²,

Hanaoka³

2019

238

113

View full text Add to dashboard Cite

show abstract

“…In general terms, as age increased, performance of practitioners improved, a finding consistent with anecdotes provided by facial comparison practitioners [2, 4] and results from previous facial comparison [13] and commercial facial recognition algorithm evaluations [2, 12, 21]. Even when the age variation was kept constant at 0 years, performance was worse for younger ages in childhood.…”

Section: Discussionsupporting

confidence: 77%

“…Although many government agencies conduct facial image comparisons of children with age variations often ranging from 0–10 years [2], research that has examined performance in this area is scant. Only one previous study has employed a comprehensive evaluation across age in childhood using operational images, however this study evaluated the performance of commercial facial recognition algorithms, not practitioners [12].…”

Section: Introductionmentioning

confidence: 99%

The performance of practitioners conducting facial comparisons on images of children across age

2019

Self Cite

View full text Add to dashboard Cite

Determining the identity of children is critical to aid in the fight against child exploitation, as well as for passport control and visa issuance purposes. Facial image comparison is one method that may be used to determine identity. Due to the substantial amount of facial growth that occurs in childhood, it is critical to understand facial image comparison performance across both chronological age (the age of the child), and age variation (the age difference between images). In this study we examined the performance of 120 facial comparison practitioners from a government agency on a dataset of 23,760 image pairs selected from the agency’s own database of controlled, operational images. Each chronological age in childhood (0–17 years) and age variations ranging from 0–10 years were examined. Practitioner performance was found to vary considerably across childhood, and depended on whether the pairs were mated (same child) or non-mated (different child). Overall, practitioners were more accurate and confident with image pairs containing older children, and also more accurate and confident with smaller age variations. Chronological age impacted on accuracy with mated pairs, but age variation did not. In contrast, both age and age variation impacted on accuracy with non-mated pairs. These differences in performance show that changes in the face throughout childhood have a significant impact on practitioner performance. We propose that improvements in accuracy may be achievable with a better understanding of which facial features are most appropriate to compare across childhood, and adjusting training and development programs accordingly.

show abstract

“…They observed that ICAO SDKs work better with Caucasian subjects than with African Americans. The work from [20] made an extensive study analyzing several age cohorts using one COTS system. Among several observations made, the most impacting one was the high FMR and high False Non-Match Rates (FNMR) in pairs of images where age is lower than four years old.…”

Section: Fairness In Biometricsmentioning

confidence: 99%

Fairness in Biometrics: A Figure of Merit to Assess Biometric Verification Systems

Pereira

Marcel

2022

IEEE Trans. Biom. Behav. Identity Sci.

View full text Add to dashboard Cite

Machine learning-based (ML) systems are being largely deployed since the last decade in a myriad of scenarios impacting several instances in our daily lives. With this vast sort of applications, aspects of fairness start to rise in the spotlight due to the social impact that this can get in some social groups. In this work aspects of fairness in biometrics are addressed. First, we introduce a figure of merit that is able to evaluate and compare fairness aspects between multiple biometric verification systems, the so-called Fairness Discrepancy Rate (FDR). A use case with two synthetic biometric systems is introduced and demonstrates the potential of this figure of merit in extreme cases of demographic differentials. Second, a use case using face biometrics is presented where several systems are evaluated compared with this new figure of merit using three public datasets exploring gender and race demographics.

show abstract

The Impact of Age and Threshold Variation on Facial Recognition Algorithm Performance Using Images of Children

Cited by 23 publications

References 14 publications

Face recognition vendor test part 3:

Face recognition vendor test part 3:

The performance of practitioners conducting facial comparisons on images of children across age

Fairness in Biometrics: A Figure of Merit to Assess Biometric Verification Systems

Contact Info

Product

Resources

About