In This Story
When people ask George Mason University Associate Professor David Lattanzi why a civil engineer is working in forensic nursing science, he can only say, ‘Why not?’ Lattanzi is a co-principal investigator on research to develop a tool that will help clinicians and others assess bruises on victims of violence, particularly those with darker skin tones.
This work, led by principal investigator Katherine Scafide, associate professor in the College of Public Health, recently received a $4.86 million gift from an anonymous donor which extends the findings of an earlier Department of Justice grant that began more than three years ago.
The team’s multidisciplinary research combines Scafide’s work with alternate light source technology, co-PI Janusz Wojtusiak’s expertise in informatics, and Lattanzi’s knowledge of computer vision and deep learning to address challenges of identifying bruises and other injuries in victims of domestic violence.
“When I started at Mason 10 years ago, I never imagined that my work could have a societal impact like this,” said Lattanzi. But Lattanzi has always been interested in finding innovative solutions to real-world problems. A structural engineer by training, he spent his early years as a bridge and tunnel inspector. After seeing too many close calls while working on inspection sites, he thought there must be a better way. So he began using drones and imaging to address the next generation of infrastructure-inspection technologies—and make bridge inspection less dangerous.
Though drones aren’t part of this research, he and the team are using the image-based diagnostic techniques that began a decade ago. Now Lattanzi is enhancing these tools with artificial intelligence (AI) and machine learning to examine human skin structures damaged by incidents of domestic violence.
According to a statement from the university, one in three women worldwide experience some form of violence. In the United States, more than 10 million women and men deal with interpersonal violence each year. These numbers tell only part of the story, because bruises are difficult to detect on darker skin tones, many victims miss out on getting the help they need.
The tool these researchers will build needs a large enough database to identify skin tones “accurately and equitably,” said Lattanzi. One of the important goals of the project, therefore, is to build a unique data repository that combines images of bruises and other injuries, with measurements, clinical and demographic information about the victim, and information inferred by artificial intelligence. Currently, the platform includes about 30,000 images of bruises collected using visible and alternative light sources.
“Having a massive database leads to better outcomes and measurably better outcomes for those victims downstream in terms of their care and in terms of their outcomes in the criminal justice system,” Lattanzi said.
The tool the Mason researchers are building could be an app on a smartphone that clinicians can use to assess the age of bruises. Lattanzi said they hope to have a prototype in a year, and possibly have the technology available for use in a couple of years.