Computer scientists at UNC-Chapel Hill and colleagues at the Swiss university, ETH-Zurich, have created an algorithm that searches through millions of photos on Flickr, then uses them to build a 3D model of landmarks and geographical locations. Case in point, the video above. According to The Daily Tar Heel, “researchers demonstrated the technique by using 3 million images of Rome to reconstruct the city’s primary landmarks. A single PC processed the images in less than 24 hours. Landmarks in Berlin were reconstructed in the same manner.” Not bad for a day’s work…
via ReadWriteWeb
Add a comment