ComÂputÂer sciÂenÂtists at UNC-Chapel Hill and colÂleagues at the Swiss uniÂverÂsiÂty, ETH-Zurich, have creÂatÂed an algoÂrithm that searchÂes through milÂlions of phoÂtos on Flickr, then uses them to build a 3D modÂel of landÂmarks and geoÂgraphÂiÂcal locaÂtions. Case in point, the video above. AccordÂing to The DaiÂly Tar Heel, “researchers demonÂstratÂed the techÂnique by using 3 milÂlion images of Rome to reconÂstruct the city’s priÂmaÂry landÂmarks. A sinÂgle PC processed the images in less than 24 hours. LandÂmarks in Berlin were reconÂstructÂed in the same manÂner.” Not bad for a day’s work…
via ReadÂWriteWeb