Age Estimation Using Convolutional Neural Networks with Lumbar and Thoracic Spine Images from Postmortem Computed Tomography: A Pilot Study
Open Access
- 28 February 2026
- journal article
- Published by The Korean Society for Legal Medicine in Korean Journal of Legal Medicine
- Vol. 50 (1) , 1-8
Abstract
In forensic medicine, age estimation commonly involves assessing age-related changes in teeth and skeletal structures. Vertebral morphological alterations, such as osteophyte formation, serve as age indicators. Recent studies using deep-learning techniques, such as neural networks, for age estimation from radiographic images have been conducted, reporting significantly higher accuracy than previous studies. This study aimed to estimate age using neural network-based deep-learning techniques applied to computed tomography (CT) cross-sectional images of the spine and evaluate its feasibility. Postmortem CT scans of 214 cadavers with varying decomposition levels were used. Coronal and sagittal cross-sectional images penetrating the center of each vertebral body were extracted for the 11th and 12th thoracic vertebrae and the first to fifth lumbar vertebrae. Using these images, along with the chronological ages of deceased individuals, an age estimation model was developed through regression analysis in PyTorch, employing a convolutional neural networks architecture with five-fold cross-validation. The model achieved a mean absolute error of 5.385 years, root mean squared error of 7.029 years, and coefficient of determination of 0.793. Although the sample size was relatively small, the results suggested the potential applicability of vertebral imagingbased age estimation in the Korean population. Further research using a larger dataset may improve the accuracy and reliability of the model.Keywords
Funding Information
- National Forensic Service (NFS2025CLI11)
- Ministry of the Interior and Safety
This publication has 23 references indexed in Scilit:
- Automatic and robust estimation of sex and chronological age from panoramic radiographs using a multi-task deep learning network: a study on a South Korean populationInternational journal of legal medicine, 2024
- Age estimation using vertebral bone spurs; Testing the efficacy of three methods on a European populationForensic Science International: Reports, 2022
- Age estimation by evaluation of osteophytes in thoracic and lumbar vertebrae using postmortem CT images in a modern Japanese populationInternational journal of legal medicine, 2021
- Development of an age estimation method for bones based on machine learning using post-mortem computed tomography images of bonesForensic Imaging, 2021
- Cranial suture closure as an age indicator: A reviewForensic Science International, 2020
- Analysis of forensic odontological examinations at the National Forensic Service of Korea from 2011 to 2015Legal Medicine, 2018
- The application of the Kvaal method to estimate the age of live Korean subjects using digital panoramic radiographsInternational journal of legal medicine, 2017
- DNA methylation-based age prediction from saliva: High age predictability by combination of 7 CpG markersForensic Science International: Genetics, 2017
- Age estimation from the degree of osteophyte formation of vertebral columns in JapaneseLegal Medicine, 2006
- Age estimation from the auricular surface of the ilium: A revised methodAmerican Journal of Physical Anthropology, 2002