Researchers from the Commonwealth Scientific and Industrial Research Organisation, along with the University of Western Australia, have created a sophisticated artificial intelligence (AI) tool to help with estimating biological sex from human skulls. The study was published in the journal Scientific Reports.
A volume-rendered CT scan featuring the five cranial traits used in the study. Image Credit: Commonwealth Scientific and Industrial Research Organisation
When results are required quickly, as in criminal investigations and major natural disasters, the AI tool could help investigators by speeding up the accurate identification of skulls.
The AI tool attained a 97% accuracy rate, significantly surpassing the 82% accuracy of conventional methods employed by human assessors.
The tool was created in partnership with The University of Western Australia (UWA), whose forensic anthropology specialists contributed domain knowledge and labeled data to aid in the creation of the model.
According to CSIRO Research Scientist and Study Co-First Author Dr. Hollie Min, the AI algorithm analyzed 200 computerized tomography (CT) scans for sex-associated characteristics and compared the results to human analysis.
Our AI tool produces its results approximately five times faster than humans can, meaning families waiting for results of investigations can receive news about their loved ones more quickly. This AI tool has the potential to support forensic anthropologists to enhance the accuracy of sex estimations while reducing the potential impact of human bias.
Dr. Hollie Min, Research Scientist, and Study Co-First Author, Commonwealth Scientific and Industrial Research Organisation
Dr. Min also emphasized the importance of considering population-specific differences in skull characteristics.
This collaborative study allowed us to address some of the perceived limitations of traditional methods and better account for diversity in forensic data. Future research is needed, especially around expanding datasets to include diverse populations, enhancing the robustness and generalisability of the AI framework.
Dr. Hollie Min, Research Scientist and Study Co-First Author, Commonwealth Scientific and Industrial Research Organisation
Min said, “Our goal is to provide forensic anthropologists with a reliable, interpretable tool to support their critical work, especially in cases involving individuals of unknown population backgrounds.”
This cooperative endeavor shows how AI can help forensic anthropology and progress the discipline with creative, data-driven solutions.
“Our team is currently looking for industry collaborators to develop and translate this technology for real-life applications,” Dr Min added.
The CT database was gathered at Hasanuddin University's Dr. Wahidin Sudirohusodo General Hospital (RSWS) in Indonesia.
Journal Reference:
Lye, R., et al. (2024) Deep learning versus human assessors: forensic sex estimation from three-dimensional computed tomography scans. Scientific Reports. doi.org/10.1038/s41598-024-81718-y.