Sex Determination from Panoramic Dental X-ray Images – Comparison of Artificial Intelligence and Human Results

Abstract

Background: Accurate sex estimation represents a fundamental step in the human identification process. It is frequently performed by trained experts who measure mandibular parameters and evaluate results to differentiate male from female human remains. With the growth of digital sciences, AI can now provide fast and precise results in many fields, including forensic odontology. It was first conceptualized in the 1950s, and since then, its usage has been exponentially expanding.
Aim: The aim of this preliminary study was to evaluate the accuracy of artificial intelligence in sex determination in comparison with human observers.
Material and Methods: This research compares the accuracy in sex determination based on orthopantomographic digital images. Radiographs were evaluated by AI and doctors of dental medicine. Fifteen individuals (doctors of dental medicine) participated in this preliminary study. They were asked to determine patient sex based on the presented OPG images. The artificial intelligence approach was based on the application of deep learning techniques for analysis of dental X-ray images, using deep convolutional neural networks (CNNs).
Results: Previously trained deep convolutional neural networks achieved 100% accuracy in sex determination, while doctors of dental medicine achieved a mean accuracy of 74.40%.
Conclusions: Despite the limited sample size, the high accuracy observed in this preliminary study suggests that artificial intelligence algorithms may have potential as supportive tools for sex determination in forensic odontology; however, further research involving larger and more diverse datasets is needed to establish their reliability and generalisability.