Automatic Age Estimation and Majority Age Classification from Multi-Factorial MRI Data

Darko Stern, Christian Payer, Nicola Giuliani, Martin Urschler

Research output: Contribution to journalArticleResearchpeer-review

Abstract

Age estimation from radiologic data is an important topic both in clinical medicine as well as in forensic applications, where it is used to assess unknown chronological age or to discriminate minors from adults. In this work, we propose an automatic multi-factorial age estimation method based on MRI data of hand, clavicle and teeth to extend the maximal age range from up to 19 years, as commonly used for age assessment based on hand bones, to up to 25 years, when combined with clavicle bones and wisdom teeth. Fusing age-relevant information from all three anatomical sites, our method utilizes a deep convolutional neural network that is trained on a dataset of 322 subjects in the age range between 13 and 25 years, to achieve a mean absolute prediction error in regressing chronological age of 1.01 +/- 0.74 years. Furthermore, when used for majority age classification, we show that a classifier derived from thresholding our regression based predictor is better suited than a classifier directly trained with a classification loss, especially when taking into account that cases of minors being wrongly classified as adults need to be minimized. In conclusion, we overcome the limitations of the multi-factorial methods currently used in forensic practice, i.e., dependency on ionizing radiation, subjectivity in quantifying age-relevant information, and lack of an established approach to fuse this information from individual anatomical sites.

Original languageEnglish
Pages (from-to)1392-1403
JournalIEEE Journal of Biomedical and Health Informatics
Volume23
Issue number4
DOIs
Publication statusPublished - Jul 2019

Fingerprint

Magnetic resonance imaging
Minors
Clavicle
Bone
Classifiers
Ionizing radiation
Electric fuses
Hand Bones
Medicine
Third Molar
Clinical Medicine
Ionizing Radiation
Neural networks
Tooth
Hand
Bone and Bones
Datasets

Keywords

  • age estimation
  • Bones
  • convolutional neural network
  • Estimation
  • Feature extraction
  • Forensics
  • Informatics
  • information fusion
  • Magnetic resonance imaging
  • magnetic resonance imaging
  • majority age classification
  • multi-factorial

ASJC Scopus subject areas

  • Biotechnology
  • Computer Science Applications
  • Electrical and Electronic Engineering
  • Health Information Management

Fields of Expertise

  • Information, Communication & Computing

Cooperations

  • BioTechMed-Graz

Cite this

Automatic Age Estimation and Majority Age Classification from Multi-Factorial MRI Data. / Stern, Darko; Payer, Christian; Giuliani, Nicola; Urschler, Martin.

In: IEEE Journal of Biomedical and Health Informatics, Vol. 23, No. 4, 07.2019, p. 1392-1403.

Research output: Contribution to journalArticleResearchpeer-review

Stern, Darko ; Payer, Christian ; Giuliani, Nicola ; Urschler, Martin. / Automatic Age Estimation and Majority Age Classification from Multi-Factorial MRI Data. In: IEEE Journal of Biomedical and Health Informatics. 2019 ; Vol. 23, No. 4. pp. 1392-1403.
@article{13f43955113241a58a4d664d1121fa88,
title = "Automatic Age Estimation and Majority Age Classification from Multi-Factorial MRI Data",
abstract = "Age estimation from radiologic data is an important topic both in clinical medicine as well as in forensic applications, where it is used to assess unknown chronological age or to discriminate minors from adults. In this work, we propose an automatic multi-factorial age estimation method based on MRI data of hand, clavicle and teeth to extend the maximal age range from up to 19 years, as commonly used for age assessment based on hand bones, to up to 25 years, when combined with clavicle bones and wisdom teeth. Fusing age-relevant information from all three anatomical sites, our method utilizes a deep convolutional neural network that is trained on a dataset of 322 subjects in the age range between 13 and 25 years, to achieve a mean absolute prediction error in regressing chronological age of 1.01 +/- 0.74 years. Furthermore, when used for majority age classification, we show that a classifier derived from thresholding our regression based predictor is better suited than a classifier directly trained with a classification loss, especially when taking into account that cases of minors being wrongly classified as adults need to be minimized. In conclusion, we overcome the limitations of the multi-factorial methods currently used in forensic practice, i.e., dependency on ionizing radiation, subjectivity in quantifying age-relevant information, and lack of an established approach to fuse this information from individual anatomical sites.",
keywords = "age estimation, Bones, convolutional neural network, Estimation, Feature extraction, Forensics, Informatics, information fusion, Magnetic resonance imaging, magnetic resonance imaging, majority age classification, multi-factorial",
author = "Darko Stern and Christian Payer and Nicola Giuliani and Martin Urschler",
year = "2019",
month = "7",
doi = "10.1109/JBHI.2018.2869606",
language = "English",
volume = "23",
pages = "1392--1403",
journal = "IEEE Journal of Biomedical and Health Informatics",
issn = "2168-2194",
publisher = "Institute of Electrical and Electronics Engineers",
number = "4",

}

TY - JOUR

T1 - Automatic Age Estimation and Majority Age Classification from Multi-Factorial MRI Data

AU - Stern, Darko

AU - Payer, Christian

AU - Giuliani, Nicola

AU - Urschler, Martin

PY - 2019/7

Y1 - 2019/7

N2 - Age estimation from radiologic data is an important topic both in clinical medicine as well as in forensic applications, where it is used to assess unknown chronological age or to discriminate minors from adults. In this work, we propose an automatic multi-factorial age estimation method based on MRI data of hand, clavicle and teeth to extend the maximal age range from up to 19 years, as commonly used for age assessment based on hand bones, to up to 25 years, when combined with clavicle bones and wisdom teeth. Fusing age-relevant information from all three anatomical sites, our method utilizes a deep convolutional neural network that is trained on a dataset of 322 subjects in the age range between 13 and 25 years, to achieve a mean absolute prediction error in regressing chronological age of 1.01 +/- 0.74 years. Furthermore, when used for majority age classification, we show that a classifier derived from thresholding our regression based predictor is better suited than a classifier directly trained with a classification loss, especially when taking into account that cases of minors being wrongly classified as adults need to be minimized. In conclusion, we overcome the limitations of the multi-factorial methods currently used in forensic practice, i.e., dependency on ionizing radiation, subjectivity in quantifying age-relevant information, and lack of an established approach to fuse this information from individual anatomical sites.

AB - Age estimation from radiologic data is an important topic both in clinical medicine as well as in forensic applications, where it is used to assess unknown chronological age or to discriminate minors from adults. In this work, we propose an automatic multi-factorial age estimation method based on MRI data of hand, clavicle and teeth to extend the maximal age range from up to 19 years, as commonly used for age assessment based on hand bones, to up to 25 years, when combined with clavicle bones and wisdom teeth. Fusing age-relevant information from all three anatomical sites, our method utilizes a deep convolutional neural network that is trained on a dataset of 322 subjects in the age range between 13 and 25 years, to achieve a mean absolute prediction error in regressing chronological age of 1.01 +/- 0.74 years. Furthermore, when used for majority age classification, we show that a classifier derived from thresholding our regression based predictor is better suited than a classifier directly trained with a classification loss, especially when taking into account that cases of minors being wrongly classified as adults need to be minimized. In conclusion, we overcome the limitations of the multi-factorial methods currently used in forensic practice, i.e., dependency on ionizing radiation, subjectivity in quantifying age-relevant information, and lack of an established approach to fuse this information from individual anatomical sites.

KW - age estimation

KW - Bones

KW - convolutional neural network

KW - Estimation

KW - Feature extraction

KW - Forensics

KW - Informatics

KW - information fusion

KW - Magnetic resonance imaging

KW - magnetic resonance imaging

KW - majority age classification

KW - multi-factorial

UR - http://www.scopus.com/inward/record.url?scp=85054283148&partnerID=8YFLogxK

U2 - 10.1109/JBHI.2018.2869606

DO - 10.1109/JBHI.2018.2869606

M3 - Article

VL - 23

SP - 1392

EP - 1403

JO - IEEE Journal of Biomedical and Health Informatics

JF - IEEE Journal of Biomedical and Health Informatics

SN - 2168-2194

IS - 4

ER -