CoChat
Sign Up
Cited by 33Forensic Anthropology and Bioarchaeology StudiesFace recognition and analysisMorphological variations and asymmetry

The DOI matches its Crossref record; 63 of the 63 references it cites were found in a registry or an index; none of the 72 sources its claims cite is checked for support yet.

Simulation of facial growth based on longitudinal data: Age progression and age regression between 7 and 17 years of age using 3D surface data

Open access
Koudelová, Jana, Hoffmannová, Eva, Dupej, Ján, Velemínská, Jana·2019·PLOS ONE
DOI 10.1371/journal.pone.0212618PMC PMC6386244
Open this paper in CoChat — free accountOpen PDF

Abstract

Modelling of the development of facial morphology during childhood and adolescence is highly useful in forensic and biomedical practice. However, most studies in this area fail to capture the essence of the face as a three-dimensional structure. The main aims of our present study were (1) to construct ageing trajectories for the female and male face between 7 and 17 years of age and (2) to propose a three-dimensional age progression (age -regression) system focused on real growth-related facial changes. Our approach was based on an assessment of a total of 522 three-dimensional (3D) facial scans of Czech children (39 boys, 48 girls) that were longitudinally studied between the ages of 7 to 12 and 12 to 17 years. Facial surface scans were obtained using a Vectra-3D scanner and evaluated using geometric morphometric methods (CPD-DCA, PCA, Hotelling's T2 tests). We observed very similar growth rates between 7 and 10 years in both sexes, followed by an increase in growth velocity in both sexes, with maxima between 11 and 12 years in girls and 11 to 13 years in boys, which are connected with the different timing of the onset of puberty. Based on these partly different ageing trajectories for girls and boys, we simulated the effects of age progression (age regression) on facial scans. In girls, the mean error was 1.81 mm at 12 years and 1.7 mm at 17 years. In boys, the prediction system was slightly less successful: 2.0 mm at 12 years and 1.94 mm at 17 years. The areas with the greatest deviations between predicted and real facial morphology were not important for facial recognition. Changes of body mass index percentiles in children throughout the observation period had no significant influence on the accuracy of the age progression models for both sexes.

Overview

AI snapshot

Written from the paper’s full text.

TL;DR

This longitudinal 3D study models facial growth in Czech children aged 7–17 and builds sex-specific age progression/regression models. It finds puberty-related growth peaks and prediction errors generally under 2 mm, with BMI percentile changes not significantly affecting accuracy.

Key findings

  • Growth rates were very similar between 7 and 10 years in both sexes, then increased, with maxima between 11 and 12 years in girls and 11 to 13 years in boys, connected to different puberty timing.
  • Age progression/regression mean error was 1.81 mm at 12 years and 1.7 mm at 17 years in girls; in boys it was 2.0 mm at 12 years and 1.94 mm at 17 years. Largest deviations (>3 mm) occurred only in marginal parts of the face.
  • The algorithm consistently improved prediction over using a non-aged face (p<0.001 in all cases); in males prediction improved on average by 2.7 mm and 2.0 mm in 7-year-olds and 12-year-olds, and in females by 3.0 mm and 1.1 mm, respectively.
  • Changes in BMI percentiles over the observation period had no significant effect on mean error (boys p = 0.925; girls p = 0.0804).

Methods

  • Longitudinal sample of 522 3D facial scans from Czech children (39 boys, 48 girls), studied in two sets scanned annually from 7–12 and 12–17 years; scans obtained with a Vectra-3D scanner, manually processed, and simplified to approx. 25k vertices.
  • Geometric morphometrics used CPD-DCA for homologous surface representations with nine landmarks, GPA, PCA, and Hotelling’s T2 tests; ageing trajectories were modelled on principal component scores using linear models per longitudinal group and Hermitian cubic interpolation.
  • Prediction was tested by simulating ageing from each specimen’s youngest to oldest known occurrence and comparing vertex distances to the real face; paired t-tests compared errors against non-aged prediction, and ANOVA tested BMI percentile effects.

Limitations

  • Stated: the framework assumes each sex exhibits a common ageing pattern and does not take into account initial form, changes in body composition, or other factors that could improve predictions; all predictions should be considered tentative.
  • Stated: BMI does not distinguish body fat from muscle mass, though body fat must be considered in studies of facial soft tissues; future work should include facial texture to improve age progression/regression models.
  • Evident: the sample is Czech, Caucasian, and drawn from specific schools, so the population-specific ontogenetic trajectories may not generalize; the 3D surface-only models omit texture and other recognition-relevant cues.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1371/journal.pone.0212618

Cites

63 of 63 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Verification

Claims the paper makes

0 of 72 cited sources checked · 0 supported · 0 flagged

  • Findingown data

    In girls, the most evident growth changes occurred between 7 and 13 years of age.

    “In girls (Fig 1) the most evident growth changes occurred between 7 and 13 years of age.”
    Quote in the paperNot a substantial claim
  • Findingown data

    In boys, between 7 and 11 years of age, slight growth of the forehead, the nose, the lips area and the chin was observed.

    “In boys (Fig 2) between 7 and 11 years of age, slight growth of the forehead, the nose, the lips area and the chin was observed.”
    Quote in the paperNot a substantial claim
  • Findingown data

    In boys, the global ageing trajectory showed the same annual changes up until 10 years of age.

    “The global ageing trajectory showed the same annual changes up until 10 years of age.”
    Quote in the paperNot a substantial claim
    • The authors state this as an assumption: The ageing model in each longitudinal group can be simplified to a linear one.
    • The authors state this as an assumption: The two partial trajectories (7–12 and 12–17 years) can be joined by a Hermitian cubic function that smoothly interpolates between their centres.
  • Findingown data

    In girls, noticeable growth changes in the cheeks (mostly their lateral parts) were apparent up until the age of 14 years.

    “In the area of the cheeks (mostly their lateral parts), noticeable growth changes were apparent up until the age of 14 years.”
    Quote in the paperNot a substantial claim
  • Findingown data

    The most evident growth changes concerned the whole forehead region, the nose region, and the mandible in both sexes.

    “The most evident growth changes concerned the whole forehead region, the nose region and the mandible in both sexes.”
    Quote in the paperSubstantial claim
  • It may state two separate claims as one. 39
  • Probably not a substantial claim. 37
  • The quote only partly states this claim. 4

  • It may state two separate claims as one.

    Findingown data

    In boys, a flattening of the cheeks (especially between 13 and 15 years) and deepening of the orbit region in relation to more prominent eyebrow ridges occurred in all age categories and became more marked with age.

    “In addition, a flattening of the cheeks (especially between 13 and 15 years) and deepening of the orbit region in relation to more prominent eyebrow ridges occurred in all age categories and became more marked with age.”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Findingown data

    In girls, the most intensive and pronounced facial growth changes occurred between 10 and 14 years of age and were much more pronounced between 11 and 12 years.

    “The most intensive and pronounced facial growth changes occurred between 10 and 14 years of age and were much more pronounced between 11 and 12 years.”
    Quote in the paperSubstantial claim
    • The authors state this as an assumption: The ageing model in each longitudinal group can be simplified to a linear one.
    • The authors state this as an assumption: The two partial trajectories (7–12 and 12–17 years) can be joined by a Hermitian cubic function that smoothly interpolates between their centres.

  • It may state two separate claims as one.

    Findingown data

    In girls, after age 13 facial growth slowed down, and after age 16 it practically ceased.

    “After the age of 13 years, the facial growth slowed down, and after 16 years of age it practically ceased.”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Findingown data

    In girls, noticeable facial changes occurred between 7 and 13 years of age, with the most significant changes between 12 and 13 years of age.

    “In girls, we observed noticeable changes between 7 and 13 years of age, with the most significant changes occurring between 12 and 13 years of age.”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Findingown data

    Between 10 and 14 years of age, growth velocity increased in both sexes, with maxima between 11 and 12 years in girls and between 11 and 13 years in boys.

    “Between 10 and 14 years of age, we observed increasing of growth velocity in both sexes, with maxima between 11 and 12 years in girls and between 11 and 13 years in boys.”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Findingown data

    In girls, the region around the eyes became deeper in relation to the facial plane, except in the age categories from 14 to 16 years, where anterior growth of the region under the eyes occurred.

    “The region around the eyes became deeper in relation to the facial plane, except in the age categories from 14 to 16 years, where anterior growth of the region under the eyes occurred.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Findingown data

    In girls, the slope of the regression line indicates that the magnitude of changes in BMI percentiles was inversely proportional to the mean error of the age-progression (regression) model.

    “The slope of the regression line indicates that the magnitude of the changes in BMI percentiles was inversely proportional to the mean error of the age-progression (regression) model.”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Findingown data

    In females, the prediction was improved on average by 3.0 mm in 7-year-olds and 1.1 mm in 12-year-olds.

    “Similarly, in females the prediction was improved by 3.0 mm and 1.1 mm, respectively.”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Findingown data

    In girls, an increasing prominence of the eyebrow ridges, nose tip and chin was observed from 7 to 14 years of age, with the greatest accentuation between the ages of 10 and 14 years.

    “An increasing prominence of the eyebrow ridges, nose tip and chin was observed from 7 to 14 years of age, with the greatest accentuation between the ages of 10 and 14 years; after this age, no marked changes in these parts were apparent.”
    Quote in the paperNot a substantial claim

  • The quote only partly states this claim.

    Findingown data

    In girls, after the age of 16 years, growth practically stopped.

    “and after the age of 16 years, growth practically stopped.”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Findingown data

    In girls, the mean error of the age progression simulation was 1.81 mm at 12 years and 1.7 mm at 17 years.

    “In girls, the mean error was 1.81 mm at 12 years and 1.7 mm at 17 years.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: An individual's facial ageing can be simulated by adding the average ageing trajectory in principal component space to that individual's scores.

  • It may state two separate claims as one.

    Findingown data

    In boys, facial changes were evident over the entire observation period up to age 17, but after age 16 growth changes were observed only in a minor part of the nose and chin.

    “Contrary to girls, facial changes in boys were evident over the entire observation period (up to the age of 17 years), but after the age of 16 years, growth changes were observed only in a minor part of the nose and chin.”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Findingown data

    In boys, the most noticeable growth changes occurred after 11 years of age and continued until the age of 14, especially in terms of anterior growth of the forehead and eyebrow ridges, increase of nose prominence and enlargement of the downward projection of the chin.

    “The most noticeable growth changes occurred after 11 years of age and continued until the age of 14, especially in terms of anterior growth of the forehead and eyebrow ridges, increase of nose prominence and enlargement of the downward projection of the chin.”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Findingown data

    In boys, PC2 explained 8.5% of total variability and was associated with facial concavity/convexity.

    “PC2 explained 8.5% of total variability and was associated with facial concavity/convexity.”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Findingown data

    In boys, the most significant growth occurred from age 10 to 14 years, with the maximum between 11 and 13 years of age.

    “The most significant growth occurred from the age 10 to that of 14 years, with the maximum between 11 and 13 years of age.”
    Quote in the paperSubstantial claim
    • The authors state this as an assumption: The ageing model in each longitudinal group can be simplified to a linear one.
    • The authors state this as an assumption: The two partial trajectories (7–12 and 12–17 years) can be joined by a Hermitian cubic function that smoothly interpolates between their centres.

  • It may state two separate claims as one.

    Findingown data

    In girls, the mean error between real and predicted facial morphology was 1.81 mm at 12 years of age and 1.7 mm at 17 years of age.

    “In girls, the mean error at 12 years of age was 1.81 mm, and at 17 years it reached 1.7 mm.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Findingown data

    According to the slope of the linear regression line and the distribution of individuals, some differences between the two sexes were observed in the relationship between BMI percentile changes and mean error.

    “Nevertheless, according to the slope of the linear regression line and the distribution of the individuals (Fig 7), some differences between the two sexes were observed.”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Findingreasoning

    Facial changes were more intensive in boys compared to girls in most age categories except 10-11 years, in which changes were most noticeable in girls.

    “Generally, facial changes were more intensive in boys compared to girls in most of the age categories except for the category between 10 and 11 years, in which the changes were the most noticeable in girls.”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Findingown data

    In girls, PC2 explained 6.4% of total variability and was associated with changes affected by facial concavity/convexity.

    “PC2 explained 6.4% of total variability and was associated with changes affected by facial concavity/convexity.”
    Quote in the paperNot a substantial claim

  • The quote only partly states this claim.

    Findingown data

    The present study indicates a slight influence of BMI percentile changes on the accuracy of the model in girls.

    “Nevertheless, our present study indicates a slight influence on the accuracy of the model in girls.”
    Quote in the paperNot a substantial claim

  • The quote only partly states this claim.

    Findingown data

    In boys, the first principal component (PC1) explained 79.2% of total variability.

    “The first principal component (PC1) explained 79.2% of total variability and represents the facial ageing process.”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Findingown data

    In boys, the mean error between real and predicted faces was 2.0 mm at 12 years of age and 1.94 mm at 17 years of age.

    “In boys, the mean error between real and predicted faces at 12 years of age was 2.0 mm, and at 17 years of age it reached 1.94 mm.”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Findingown data

    In boys, after 14 years of age, facial growth slowly decreased, but still persisted up until the end of the observation period.

    “After 14 years of age, facial growth slowly decreased, but still persisted up until the end of the observation period.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Study design and conductnone given

    The study was approved by the Institutional Review Board of Charles University, Faculty of Science on September 2011.

    “The study was approved by the Institutional Review Board of Charles University, Faculty of Science on September 2011.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Premisecitation

    The optimal tool for measuring changes in adiposity during childhood is the body mass index (BMI), either by itself or as the proportional (percentage) change in BMI, which can be adjusted for sex and age.

    “The optimal tool for measuring changes in adiposity during childhood is the body mass index (BMI), either by itself or as the proportional (percentage) change in BMI, which can be adjusted for sex and age [33].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [33] TJ Cole, 2005, What is the best measure of adiposity change in growing children: BMI, BMI %, BMI z-score or BMI centile?, Eur J Clin Nutr source

  • The quote only partly states this claim.

    Interpretationown data

    In boys, PC1 represents the facial ageing process.

    “The first principal component (PC1) explained 79.2% of total variability and represents the facial ageing process.”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Premisecitation

    Facial morphology changes with time and does not progress at a uniform pace.

    “Facial morphology changes with time and does not progress at a uniform pace [27].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [27] TJ Hutton, 2003, Estimating average growth trajectories in shape-space using kernel smoothing, IEEE Trans Med Imaging source

  • Probably not a substantial claim. It may state two separate claims as one.

    Premisecitation

    Our previous study dealt with longitudinal facial growth of juveniles aged between 12 and 15 years, was focused on ontogenetic trajectories and proposed an age progression model.

    “Our previous study [23], which dealt with longitudinal facial growth of juveniles aged between 12 and 15 years, was focused on ontogenetic trajectories and proposed an age progression model.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [23] J Koudelová, 2015, Modelling of facial growth in Czech children based on longitudinal data: Age progression from 12 to 15 years using 3D surface models, Forensic Sci Int source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Premisecitation

    Puberty causes significant changes to the face over a short period of time.

    “For example, puberty causes significant changes to the face over a short period of time [28].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [28] DH Enlow, 1996, Essentials of facial growth source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Premisecitation

    Primozic et al. [60] reported that the amount and velocity of facial growth in both sexes appears similar throughout 6-12 years of age, irrespective of the presence of a pubertal growth spurt.

    “According to their study, the amount and velocity of facial growth in both sexes appears to be similar throughout the period of 6–12 years of age, irrespective of the presence of a pubertal growth spurt.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [60] J Primozic, 2017, Facial soft tissue changes during the pre-pubertal and pubertal growth phase: a mixed longitudinal laser-scanning study, Eur J Orthod source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Premisecitation

    Bulygina et al. [51] found that male and female growth trajectories are not always parallel but undergo some degree of divergence after age about 12 years.

    “Bulygina et al. [51] similarly found that male and female growth trajectories are not always parallel but undergo some degree of divergence after the age of about 12 years.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [51] E Bulygina, 2006, Ontogeny of facial dimorphism and patterns of individual development within one human population, Am J Phys Anthropol source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

    Premisecitation

    In a previous study [23], we reported differences in facial morphology and divergences in growth trajectories between the two sexes over the observation period.

    “We reported differences in facial morphology as well as divergences in growth trajectories between the two sexes over the observation period.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [23] J Koudelová, 2015, Modelling of facial growth in Czech children based on longitudinal data: Age progression from 12 to 15 years using 3D surface models, Forensic Sci Int source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Study design and conductnone given

    Our framework does not take into account the initial form, changes in body composition, or any other factors that would likely contribute to even more accurate predictions.

    “Our framework does not take into account the initial form, changes in body composition or any other factors that would likely contribute to even more accurate predictions.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim. It may be too vague to check against evidence. It may state two separate claims as one.

    Premisecitation

    Age progression images are difficult to get right, subjective and variable even when a high degree of anatomical and artistic modelling expertise is applied.

    “Nevertheless, the they are difficult to get right, subjective and variable even when a high degree of anatomical and artistic modelling expertise is applied [20].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [20] JM Lampinen, 2015, Mighty Morphin ‘ age progression: how artist, age range, and morphing influences the similarity of forensic age progressions to target individuals, Psychol Crime Law source

  • Probably not a substantial claim. It may be too vague to check against evidence.

    Interpretationreasoning

    Specification of population-specific ontogenetic trajectories is required.

    “Nevertheless, the specification of population-specific ontogenetic trajectories is required.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

    Premisecitation

    Developmental shape changes are less consistent in early ontogeny compared to later ontogeny, during which developmental shape changes are very similar among different regional groups.

    “Developmental shape changes are less consistent in early ontogeny compared to later ontogeny, during which developmental shape changes are very similar among different regional groups [29].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [29] SE Freidline, 2015, Ontogenetic and static allometry in the human face: Contrasting Khoisan and Inuit, Am J Phys Anthropol source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Premisecitation

    Most authors have used face ageing datasets consisting of photographs.

    “Most authors have used face ageing datasets consisting of photographs, for example AGFW, CACD, FG-Net or MORPH [14–17].”
    Quote in the paperNot a substantial claim0 of 4 sources checked
    • cites [14] X Shu, 2018, Personalized Age Progression with Aging Dictionary, IEEE Trans Pattern Anal Mach Intell source
    • cites [15] S Liu, 2017, Face Aging with Contextual Generative Adversarial Nets, Proceedings of the 2017 ACM on Multimedia Conference source
    • cites [16] CN Duong, 2017, Temporal Non-Volume Preserving Approach to Facial Age-Progression and Age-Invariant Face Recognition, IEEE I Conf Comp Vis source
    • cites [17] H Yang, 2016, Face Aging Effect Simulation using Hidden Factor Analysis Joint Sparse Representation, IEEE T Image Process source

  • Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

    Interpretationnone given

    No ageing algorithm, regardless of what parameters it accounts for, can predict the true appearance of a subject with absolute accuracy.

    “Nevertheless, no ageing algorithm, regardless of what parameters it accounts for, can predict the true appearance of a subject with absolute accuracy.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

    Premisecitation

    Bulygina et al. [51] reported a significant decline in the rate of growth at approximately 13 years of age and a cessation of growth at about 15 years of age.

    “Bulygina et al. [51] also reported a significant decline in the rate of growth at approximately 13 years of age and a cessation of growth at about 15 years of age.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [51] E Bulygina, 2006, Ontogeny of facial dimorphism and patterns of individual development within one human population, Am J Phys Anthropol source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Premisecitation

    In a previous study [23], we found a strong influence of age on facial soft tissue morphology between 12 and 15 years of age.

    “In a previous study [23] we found a strong influence of age on facial soft tissue morphology between 12 and 15 years of age.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [23] J Koudelová, 2015, Modelling of facial growth in Czech children based on longitudinal data: Age progression from 12 to 15 years using 3D surface models, Forensic Sci Int source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Interpretationnone given

    The model can be used not only for forward prediction but also retrospectively.

    “Note that this model can be used not only for forward prediction, as we have done, but also retrospectively.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: An individual's facial ageing can be simulated by adding the average ageing trajectory in principal component space to that individual's scores.

  • Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

    Premisecitation

    Matthews et al. [59] reported different peaks in the rate of growth, with girls' growth accelerating from approximately 11 years and boys' from approximately age twelve.

    “Similarly, Matthews et al. [59] reported different peaks in the rate of growth, with girls’ growth accelerating from approximately 11 years of age and boys’ accelerating from approximately age twelve.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [59] HS Matthews, 2018, Modelling 3D craniofacial growth trajectories for population comparison and classification illustrated using sex-differences, Sci Rep source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Premisecitation

    There are two main approaches to studying age progression: prototyping-based age progression and modelling-based age progression.

    “There are two main approaches to studying age progression: prototyping-based age progression and modelling-based age progression [14].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [14] X Shu, 2018, Personalized Age Progression with Aging Dictionary, IEEE Trans Pattern Anal Mach Intell source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Premisecitation

    Progress in computer science and improvements to medical imaging technologies have led to the development of computer systems for forensic facial reconstruction or for three-dimensional modelling of facial ageing using longitudinal and cross-sectional data.

    “Progress in computer science and improvements to medical imaging technologies have led to the development of computer systems for forensic facial reconstruction [21,22] or for three-dimensional modelling of facial ageing using longitudinal [23,24] and cross-sectional data [25].”
    Quote in the paperNot a substantial claim0 of 5 sources checked
    • cites [21] MA Green, 1999, Interpolating Between Computerized Three-Dimensional Forensic Facial Simulations, J Forensic Sci source
    • cites [22] MP Evison, 1999, Presenting Three-Dimensional Forensic Facial Simulations on the Internet Using VRML, J Forensic Sci source
    • cites [23] J Koudelová, 2015, Modelling of facial growth in Czech children based on longitudinal data: Age progression from 12 to 15 years using 3D surface models, Forensic Sci Int source
    • cites [24] M Mydlová, 2015, Sexual dimorphism of facial appearance in ageing human adults: A cross-sectional study, Forensic Sci Int source
    • cites [25] H Matthews, 2018, Estimating age and synthesising growth in children and adolescents using 3D facial prototypes, Forensic Sci Int source

  • It may state two separate claims as one.

    Interpretationcitation

    The divergence in facial growth patterns was closely connected with the onset of puberty, which occurs about two years earlier in girls.

    “The divergence in facial growth patterns was closely connected with the onset of puberty, which occurs about two years earlier in girls [28,51].”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [28] DH Enlow, 1996, Essentials of facial growth source
    • cites [51] E Bulygina, 2006, Ontogeny of facial dimorphism and patterns of individual development within one human population, Am J Phys Anthropol source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

    Premisecitation

    Facial shape changes associated with postnatal ontogenetic allometry are very similar among diverse regional groups, but subtle differences are present.

    “Facial shape changes associated with postnatal ontogenetic allometry are very similar among diverse regional groups, but subtle differences are present.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [29] SE Freidline, 2015, Ontogenetic and static allometry in the human face: Contrasting Khoisan and Inuit, Am J Phys Anthropol source

  • It may state two separate claims as one.

    Premisecitation

    Children change dramatically as they age, making their identification difficult.

    “Resolving cases where a child is missing for that length of time is complicated by the fact that children change dramatically as they age, making their identification difficult [10].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [10] E Cunha, 2009, The problem of aging human remains and living individuals: A review, Forensic Sci Int source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

    Premisecitation

    Relative widening of the mid- and lower face, reduction of eye height, widening of the nose, thickening of the lips and down-turning of the corners of the mouth occur with increasing body fat.

    “Generally, relative widening of the mid- and lower face, reduction of eye height, widening of the nose, thickening of the lips and down-turning of the corners of the mouth occur with increasing body fat [30–32].”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites [30] S Windhager, 2013, Body fat and facial shape are correlated in female adolescents, Am J Hum Biol source
    • cites [31] VF Ferrario, 2004, Soft Tissue Facial Morphology in Obese Adolescents: A Three-Dimensional Noninvasive Assessment, Angle Orthod
    • cites [32] V Coetzee, 2009, Facial Adiposity: A Cue to Health?, Perception source

  • Probably not a substantial claim. It may state two separate claims as one.

    Study design and conductown data

    Prediction error values were calculated for each specimen using the ageing algorithm and using no prediction at all, and compared using a paired t-test to ascertain systematic improvement.

    “To determine whether our age progression framework provided useful data, we calculated prediction error values for each specimen using the ageing algorithm and using no prediction at all, using a non-aged face for the estimation. We compared the errors using a paired t-test to ascertain if there was a systematic improvement of the prediction.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

    Premisecitation

    In Bulygina et al.'s study, the divergence mainly concerns the change of shape; the magnitude of shape change per size change seems similar between the sexes.

    “This divergence mainly concerns the change of shape; the magnitude of shape change per size change seems to be similar between the sexes in their study.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [51] E Bulygina, 2006, Ontogeny of facial dimorphism and patterns of individual development within one human population, Am J Phys Anthropol source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Premisecitation

    Most forensic artists use photographs of biological relatives (e.g. parents, siblings, grandparents) to improve the prediction of the current facial appearance.

    “Most forensic artists use photographs of biological relatives (e.g. parents, siblings, grandparents) to improve the prediction of the current facial appearance [12,13].”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [12] SD Charman, 2012, Age-progressed images may harm recognition of missing children by increasing the number of plausible targets, J Appl Res Mem Cogn source
    • cites [13] SJ Gibson, 2009, Computer assisted age progression, Forensic Sci Med Pathol source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Premisecitation

    Knowledge of age-stable features (i.e. moles, diameter of the iris) improves age progression images.

    “Knowledge of age-stable features (i.e. moles, diameter of the iris) is another factor that improves age progression images [18,19].”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [18] CEP Machado, 2017, A new approach for the analysis of facial growth and age estimation: Iris ratio, PLoS One source
    • cites [19] Z Caplova, 2017, Recognition of children on age-different images: Facial morphology and age-stable features, Sci Justice source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

    Premisecitation

    The proportion of body fat influences facial morphology and plays an important role in facial shape variation of adolescents as well as adults.

    “The proportion of body fat is another factor that influences facial morphology and plays an important role in facial shape variation of adolescents as well as adults.”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites [30] S Windhager, 2013, Body fat and facial shape are correlated in female adolescents, Am J Hum Biol source
    • cites [31] VF Ferrario, 2004, Soft Tissue Facial Morphology in Obese Adolescents: A Three-Dimensional Noninvasive Assessment, Angle Orthod
    • cites [32] V Coetzee, 2009, Facial Adiposity: A Cue to Health?, Perception source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

    Premisecitation

    BMI is generally used as an accurate tool for estimating body fat mass in obese adolescents, but differences in this index between thinner children can be largely attributable to fat-free mass, as BMI does not distinguish between body fat and muscle mass.

    “Although BMI is generally used as an accurate tool for estimating body fat mass in obese adolescents, differences in this index between thinner children can be largely attributable to fat-free mass [65], as BMI does not distinguish between body fat and muscle mass, which weighs more than fat.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [65] DS Freedman, 2005, Relation of BMI to fat and fat-free mass among children and adolescents, Int J Obes source

  • It may state two separate claims as one.

    Premisecitation

    In forensic practice, most age progressions are still produced by specialized artists, and the reliability of age progressions for the same individuals created by different artists can vary substantially.

    “In forensic practice, most age progressions are still produced by specialized artists, and the reliability of age progressions for the same individuals created by different artists can vary substantially [50].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [50] WB Erickson, 2017, When age-progressed images are unreliable: The roles of external features and age range, Sci Justice source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Premisecitation

    Postnatal divergence of trajectories is an important aspect in developing adult shape differences.

    “Generally, the postnatal divergence of trajectories is an important aspect in developing adult shape differences [61].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [61] P Mitteroecker, 2013, A brief review of shape, form, and allometry in geometric morphometrics, with applications to human facial morphology, Hystrix It J Mamm

  • Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

    Premisecitation

    External shape can affect identification accuracy and may modify the representation of internal facial features in face recognition.

    “Conversely, the external shape, which can conspicuously change during maturation, with weight gain or loss, or with changes in hairstyle, can affect identification accuracy [50] and may modify the representation of internal facial features in terms of face recognition [64].”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [50] WB Erickson, 2017, When age-progressed images are unreliable: The roles of external features and age range, Sci Justice source
    • cites [64] V Axelrod, 2010, External facial features modify the representation of internal facial features in the fusiform face area, Neuroimage source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Premisecitation

    The algorithm for facial prediction based on longitudinal 3D data was first presented in a previous study [23].

    “The algorithm for facial prediction, which was based on longitudinal 3D data, was first presented in our previous study [23] a few years ago.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [23] J Koudelová, 2015, Modelling of facial growth in Czech children based on longitudinal data: Age progression from 12 to 15 years using 3D surface models, Forensic Sci Int source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

    Premisecitation

    In a cross-sectional study [25], synthetically grown 3D faces were quantitatively compared with longitudinally collected images, and 85% of the faces were predicted correctly to within three millimetres.

    “In the cross-sectional study [25], synthetically grown 3D faces were quantitatively compared with longitudinally collected images, and 85% of the faces were predicted correctly to within three millimetres.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [25] H Matthews, 2018, Estimating age and synthesising growth in children and adolescents using 3D facial prototypes, Forensic Sci Int source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Premisecitation

    Methods of 3D face reconstruction from 2D photographs or video surveillance images have been intensively studied for many years.

    “methods of 3D face reconstruction from 2D photographs or images from video surveillance cameras have been intensively studied for many years [47–49].”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites [47] J Kittler, 2005, 3D Assisted Face Recognition: A Survey of 3D Imaging, Modelling and Recognition Approachest, IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR’05)—Workshops source
    • cites [48] S Suwajanakorn, 2014, Total Moving Face Reconstruction source
    • cites [49] H Jin, 2017, Robust 3D face modeling and reconstruction from frontal and side images, Comput Aided Geom Des source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Premisecitation

    According to Tome et al. [63], the regions with the greatest discrimination power in facial recognition are the nose and forehead.

    “According to Tome et al. [63], the regions with the greatest discrimination power in facial recognition are the nose and forehead.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [63] P Tome, 2013, Identification using face regions: Application and assessment in forensic scenarios, Forensic Sci Int source

  • It may state two separate claims as one.

    Premisecitation

    Mathematical similarity is well suited for computer vision purposes but insufficient for determining the effectiveness of age-progression techniques in the real world, which requires assessment by human subjects.

    “Although this is well suited for their purposes, it is insufficient for determining the effectiveness of age-progression techniques in the real world, which requires assessment by human subjects [12].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [12] SD Charman, 2012, Age-progressed images may harm recognition of missing children by increasing the number of plausible targets, J Appl Res Mem Cogn source

  • It may state two separate claims as one.

    Study design and conductreasoning

    The present study is based on a longitudinal evaluation, which allows direct measurement of the mean error between real and predicted faces.

    “For this reason, the present study is based on a longitudinal evaluation, which allows to directly measure the mean error between real and predicted faces.”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Premisecitation

    Investigators of computational approaches are limited by the number of high-resolution images depicting individuals over time, which are mostly two-dimensional, and by the cross-sectional nature of their approaches.

    “Investigators of computational approaches are also limited by the number of high-resolution images depicting individuals over time, which are mostly two-dimensional, and by the cross-sectional nature of their approaches [10,25].”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [10] E Cunha, 2009, The problem of aging human remains and living individuals: A review, Forensic Sci Int source
    • cites [25] H Matthews, 2018, Estimating age and synthesising growth in children and adolescents using 3D facial prototypes, Forensic Sci Int source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Premisecitation

    The shape and configuration of internal facial features (eyes, nose, and mouth) are crucial criteria in facial recognition.

    “The shape and configuration of internal facial features (i.e. the eyes, nose and mouth) are crucial criteria in facial recognition [62].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [62] CA Longmore, 2015, The importance of internal facial features in learning new faces, Q J Exp Psychol source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Premisecitation

    Numerous studies dealing with facial age progression have recently been published, mostly by computer vision scientists.

    “Numerous studies dealing with facial age progression have recently been published, mostly by computer vision scientists [13,43–46].”
    Quote in the paperNot a substantial claim0 of 5 sources checked
    • cites [13] SJ Gibson, 2009, Computer assisted age progression, Forensic Sci Med Pathol source
    • cites [43] Cardenas-Esguerra M, Vidal C, Cavalcante-Neto JB, Vieira R. Facial aging simulation applied to the missing person problem. In: Inform CLEI 2012 XXXVIII Conf Latinoam En. 2012. source
    • cites [44] X Geng, 2006, Learning from facial aging patterns for automatic age estimation, ACM Press source
    • cites [45] J Liu, 2014, Hybrid constraint SVR for facial age estimation, Signal Processing source
    • cites [46] N Ramanathan, 2006, Modeling Age Progression in Young Faces, In: Proc IEEE Conf Computer Vision and Pattern Recognition source
66 of 66

  • The authors state this as an assumption.

    Widening the age interval to 7–17 years would help identify ontogenetic traits typical for this period (e.g., onset of puberty) and improve the age progression model.

    “We supposed that the widening of the interval to earlier and later stages of facial development would help to identify certain ontogenetic traits typical for this period (i.e. the onset of puberty), which could improve the age progression model.”

  • The authors state this as an assumption.

    The ageing model in each longitudinal group can be simplified to a linear one.

    “We therefore chose to simplify the ageing model in each longitudinal group to a linear one.”

  • The authors state this as an assumption.

    The two partial trajectories (7–12 and 12–17 years) can be joined by a Hermitian cubic function that smoothly interpolates between their centres.

    “As the final step, we glued these two trajectories with a Hermitian cubic function that smoothly interpolated between their centres.”

  • The paper relies on this without support.

    An individual's facial ageing can be simulated by adding the average ageing trajectory in principal component space to that individual's scores.

    “We simulated ageing by transforming a specimen into principal component space and adding the desired portion of the ageing trajectory.”

  • The paper relies on this without support.

    The template surface selected arbitrarily is adequate for homologous representation.

    “For the next stage, a template surface (also known as the base mesh) had to be selected arbitrarily.”

  • The paper relies on this without support.

    Prediction error measured as distances between corresponding vertices is a valid measure of age progression accuracy.

    “We measured the magnitude of the error by comparing our predicted faces against the ground truth. In each age group we took the youngest occurrence of each specimen and simulated their ageing to the oldest known occurrence. That prediction was compared against the known face by calculating distances between respective vertices, which was facilitated by vertex homology.”

  • The authors state this as an assumption.

    Each sex exhibits a common ageing pattern.

    “We assume, however, that each sex exhibits a common ageing pattern.”
7 of 7

Citations

Cites

Flagged first, then by citations

Research works printed but not linked: 1

  1. VF Ferrario, 2004, Soft Tissue Facial Morphology in Obese Adolescents: A Three-Dimensional Noninvasive Assessment, Angle Orthod, 74, 37

Cited by

Sorted by citations

Related

Metadata

Authors
Koudelová, J., Hoffmannová, E., Dupej, J., & Velemínská, J.
Affiliations
Charles University (CZ)
Corresponding author
Koudelová, Jana
ORCID
Venue
PLOS ONE
Publisher
Public Library of Science (PLoS)
Published
2019-02-22
Volume
14
Issue
2
Pages
e0212618
ISSN
1932-6203 (ISSN-L)
DOI
10.1371/journal.pone.0212618
PMC
PMC6386244
OpenAlex
W2915916359
Semantic Scholar
8d7f955e6a59198fb0738d91b198382bcfe4235a
Cited by
33
References
63
Citations by year
2026: 6 · 2025: 5 · 2024: 4 · 2023: 2 · 2022: 5 · 2021: 5 · 2020: 4 · 2019: 2
FWCI
23.31
Type
journal article
Language
en
Open access
Licence
Locations
  • PLoS ONE (publishedVersion, cc-by)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
Charles University Grand Agency (656216)
Topics
  • Social Sciences › Arts and Humanities › Archeology (arts and humanities) › Forensic Anthropology and Bioarchaeology Studies
  • Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Face recognition and analysis
  • Life Sciences › Agricultural and Biological Sciences › Ecology, Evolution, Behavior and Systematics › Morphological variations and asymmetry
Keywords
facial growth, age progression, age regression, facial morphology, 3D facial scans, childhood and adolescence, sex differences, puberty onset, growth rate, face recognition, body mass index
MeSH terms
Adolescent, Child, Child Development, Face, Female, Humans, Longitudinal Studies, Male, Maxillofacial Development, Models, Biological, Sex Characteristics, Imaging, Three-Dimensional, Adolescent Development

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

A language model drew the claims from the paper on 2026-11-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-11-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review