Computer Vision News - May 2023

20 Women in Computer Vision working on the Finder project, which is this idea that I have a picture and nothing else but just the pixels in the picture. How can I figure out where in the world it was taken? So I was working with Robert on tools for the really fine-scaled refinement of camera geolocalization. Which is, if I sort of have a rough estimate of where I am, how can I really precisely define the parameters of the camera that took that picture? That was in 2012 when I started working for him. And in 2013, there was a story in our local paper about a 1980s murder victim who had been buried in a cemetery in North St Louis. The case had run cold. They didn't know who she was. And they wanted to exhume her to domodern forensic analysis, but they couldn't find her gravesite. They had pictures of the burial, though, from the 1980s. It is a testament to Robert that I've been able to have the career that I have. He gave me full support to take the things we were developing for the intelligence agencies and to instead go and help on this cold case to figure out where the 1983 picture was taken. So we went out and tried to find things like the headstones and the billboards, the things that were in the cemetery in the 1980s. If we could find where those were today, then it's a fairly straightforward optimization problem to solve for the parameters of the camera that could have taken the picture. And one of those pictures was taken right on the corner of the grave. Was this a eureka moment for you? It was, absolutely! This changed the trajectory of my career. We were able to identify her grave. As I said, the camera geometry is fairly straightforward. This was a point at which I really recognized that I was able to do something good with