News

An international research team led by Professor Yoon Yong-su of Dongseo University’s Department of Radiology has published a paper on the development of fast and accurate gender determination technology in forensic science using deep learning. The paper, “Sex Estimation Using Skull Silhouette Images from Postmortem Computed Tomography by Deep Learning” (DOI: 10.1038/s41598-024-74703-y) was published in Scientific Reports, a global journal published by Nature. Scientific Reports is a distinguished, highly ranked journal in medical engineering and science.

The study led by Professor Yoon proposes a deep learning model to determine gender based on 2D cranial silhouette images generated from post-mortem computed tomography (PMCT) data. The model offers significant advantages in terms of speed and accuracy, as it allows for quick gender estimation using simple CT scans or 2D X-ray images. Current forensic science methods for determining the gender of unidentified individuals rely mainly on visual analysis or measurement of cranial features, but these approaches are time-consuming and require considerable effort. This study addresses the limitations of these traditional methods, making it significant, particularly in large-scale disaster situations where rapid identification is crucial.

This achievement is the result of collaboration between South Korea and Japan. Professor Yoon Yong-Su and Dr. Kim Ye-Ji from Dongseo University worked with Professor Yosuke Usumoto and Professor Junji Morishita and their teams at Japan’s Kyushu University. The two Japanese teams worked in Kyushu University’s Graduate School of Medical Sciences, specializing in forensic pathology, and the same university’s Department of Quantum Medical Science and Technology. The collaboration brought together expertise from various fields, including medical artificial intelligence, health science, and forensic medicine. Based on these results, the research team has demonstrated the practicality of deep learning technology for forensic gender estimation. It is anticipated that future applications in various medical and forensic contexts could result from this research.

Additionally, the research team published a paper in the journal Diagnostics in August 2024. This paper is related to the development of a deep learning model for extracting bone information for identity verification in the field of forensic pathology. In this way, Dongseo University and Kyushu University have been continuously conducting collaborative research in the field of medical artificial intelligence, achieving significant results. Professor Yoon Yong-Su is currently conducting research projects at the National Forensic Service of Korea and continues to carry out innovative research in the fields of forensic medicine and medical artificial intelligence.