Artificial Intelligence Enhances Forensic Identification of Individuals Through Chest Radiographs

An interdisciplinary team from Michigan State University has developed an innovative application of artificial intelligence (AI) that significantly accelerates the identification process in forensic cases. By analyzing chest radiographs, the system can swiftly shortlist potential matches, vastly improving efficiency in critical situations.

The research conducted by faculty and doctoral students from the Department of Anthropology and the Department of Computer Science and Engineering has resulted in a groundbreaking study published in IEEE Access. The study scrutinized over 5,000 chest radiographs, pinpointing various regions of interest that are essential for accurately identifying individuals.

Utilizing deep neural networks, a sophisticated AI technology, the team demonstrated that their system could analyze more than 1,800 radiographs in just 17 seconds. In contrast, traditional human analysis would require between 30 to 60 hours for the same task. This remarkable speed could prove invaluable in mass fatality incidents where rapid identification is crucial.

The AI system is not only designed for immediate identification during emergencies but also extends its utility to databases of unidentified or missing persons, proposing potential matches that can help reduce bias among practitioners. The advanced algorithms compare target radiographs against thousands of others to determine the most likely identities.

This innovative approach represents a novel application of AI in forensic science, differing from the conventional medical use of radiographs, which primarily focuses on diagnosing diseases. The researchers emphasize that their work highlights the unique methodologies employed in forensic anthropology compared to medical practices.

Collaboration among researchers from various disciplines has been a key aspect of this project. The team, including experts from the iPROBE Lab in Computer Science and Engineering, has enjoyed the brainstorming sessions that combine insights from computer science and forensic anthropology, leading to a more robust understanding of the challenges and possibilities in forensic casework.

This study not only showcases the efficiency of AI in forensic applications but also illustrates the potential for future advancements in the field. By incorporating AI technologies, forensic practitioners can enhance their capabilities, providing quicker and more reliable identification methods that could save time and resources during critical investigations.