In This Story
When Associate Professor Dave Lattanzi worked as a bridge inspector, he was nearly killed three times. These close calls motivated him to find a better way to inspect infrastructure and gain additional education. “I went back to get my PhD largely to think about how we could make the [inspection] process objective and quantifiable and link it to engineering analysis,” he said. His quest led him to use drones, point cloud analysis, and digital twins to make infrastructure inspections safer and more efficient.
Since then, Lattanzi has examined structures as diverse as 20th-century bridges, ancient towers, the sunken hull of a Confederate submarine, a naval destroyer, and most recently, images of human bruises. This work produced a series of strategies and tools that promote safety, enhance resiliency, and improve health, all connected by data analysis and image diagnostics.
After his initial work with bridges, Lattanzi attended a NAVSEA Carderock research day and heard several researchers discussing a forensic analysis they were performing as they investigated the hull of a Civil War-era submarine that had been recovered from Charleston Harbor. “What’s really interesting is they had all of the same problems… the challenges map almost one to one with the challenges I saw in infrastructure inspection,” said Lattanzi.
A tragic accident around the same time connected the two when the USS McCain suffered extensive damage from a collision in the Sea of Japan. The naval researchers and Lattanzi reasoned if they could use remote sensing to map the structural damage, they could understand the ship’s capacity to get the sailors and the destroyer safely to port. They did this by using what Lattanzi calls pixels to PSI, a process of mapping the images captured at sea to the stress analysis. Over time, the Navy has adapted what Lattanzi developed at George Mason and used it in many applications including naval operations.
Most recently, Lattanzi has joined forces with researchers at George Mason University’s College of Public Health. Now, he and the team of Kat Scafide and Janusz Wojtusiak, are using what he knows about deep learning and image analysis to develop a tool to help clinicians and others assess bruises on victims of domestic violence. The research examines bruising on people with darker skin tones, who experience worse outcomes in the healthcare and legal system because their injuries are harder to see. Lattanzi hopes to develop a tool that is equitable, generalizable, and lightweight so it can be used in the field by clinicians who need it most to help vulnerable populations.
Lattanzi’s curiosity and big-picture worldview helps him translate research into practical applications. He cares about translation and emphasizes the need for engineers to develop user-friendly tools for people who need them.
“When we ask questions and think big, we can solve all kinds of problems,” he said. “Civil engineers need to dream big. We will be at the center of climate change resilience and smart cities in the future.”