Estimation of ancestry from cranial measurements based on MDCT data acquired in a Japanese and Western Australian population

Suguru Torimitsu,Akari Nakazawa,Ambika Flavel,Lauren Swift,Yohsuke Makino,Hirotaro Iwase,Daniel Franklin
DOI: https://doi.org/10.1007/s00414-024-03159-6
2024-01-23
International Journal of Legal Medicine
Abstract:The estimation of ancestry is important not only towards establishing identity but also as a required precursor to facilitating the accurate estimation of other attributes such as sex, age at death, and stature. The present study aims to analyze morphological variation in the crania of Japanese and Western Australian individuals and test predictive models based on machine learning for their potential forensic application. The Japanese and Western Australian samples comprise computed tomography (CT) scans of 230 (111 female; 119 male) and 225 adult individuals (112 female; 113 male), respectively. A total of 18 measurements were calculated, and machine learning methods (random forest modeling, RFM; support vector machine, SVM) were used to classify ancestry. The two-way unisex model achieved an overall accuracy of 93.2% for RFM and 97.1% for SVM, respectively. The four-way sex and ancestry model demonstrated an overall classification accuracy of 84.0% for RFM and 93.0% for SVM. The sex-specific models were most accurate in the female samples (♀ 95.1% for RFM and 100% for SVM; ♂91.4% for RFM and 97.4% for SVM). Our findings suggest that cranial measurements acquired in CT images can be used to accurately classify Japanese and Western Australian individuals into their respective population. This is the first study to assess the feasibility of ancestry estimation using three-dimensional CT images of the skull.
medicine, legal
What problem does this paper attempt to address?