When Katherine Scafide first started studying bruises, she noticed a frustrating problem: the tools available to detect and assess skin injuries worked great on some people but poorly on others. "Documenting injuries accurately is essential for both medical treatment and seeking justice for victims of violence," said Scafide, an associate professor of nursing at George Mason University. Now, she and her team have patented technology that could finally fix that gap.
Researchers at George Mason University's Injury Analytics Lab have created an AI system that analyzes skin images alongside patient health records to detect and assess bruises more fairly across all skin tones. The technology uses artificial intelligence combined with advanced cameras and medical data to give clinicians and forensic professionals clearer, more reliable information about injuries.
What makes this system stand out is how it adjusts to each person. Rather than using the same lighting and camera settings for everyone, the AI first measures a person's skin color and then selects the best light wavelengths and filters to capture injury evidence. For people with darker skin tones, whose injuries can be harder to see with traditional methods, this personalized approach could be especially helpful. The system can also estimate a bruise's age, size, color, and shape while explaining how confident it is in its findings.
The team includes three co-directors with different expertise: Scafide handles nursing and injury research, Janusz Wojtusiak leads the health informatics and data systems, and David Lattanzi develops the computer vision technology. This interdisciplinary approach—nurses, computer scientists, and engineers working together—helped them build a tool that actually works in real-world settings rather than just in labs.
Bruises have always been tricky to document because factors like skin tone, room lighting, and injury age can make them appear very differently from person to person. This has created real problems: injuries on darker-skinned individuals may be underdocumented or missed entirely, affecting both medical care and legal cases. The George Mason system tackles this by combining what the camera sees with information from electronic health records, including medical history, medications, and demographics. This fuller picture produces more accurate assessments than images alone could ever achieve.
The researchers say better injury documentation could improve treatment decisions, strengthen forensic evidence for investigations, and give scientists better data to develop better care practices. By making injury detection more objective and consistent, the technology has potential to reduce disparities in how injuries are evaluated across different communities. The team continues refining the system with input from the professionals who will eventually use it in hospitals, courts, and research centers.