A new computer algorithm developed at the University of Washington, using millions of online photographs from Flickr is able to construct a 3d model of various popular tourist sites. They are presently using the cities of Rome, Venice and Dubrovnik as examples to demonstrate and test the application that could allow a direct conversion of images to 3d models. The program is said to work based on its ability to calculate exact spots from where photographs were taken and then arranging various pixels to construct the 3d model (I am sure it isn't as simple as it sounds).
The 3d models somehow seem to embody the fragile translucent membrane quality, result of borrowing, selecting and careful arranging of pixels from an archive of images that are a result of a combination of eye, machine and experience, of the Colosseum, Trevi fountain, St. Peter's Basilica etc. by multiple visitors. To me it is like programming a flash mob with paint brushes carrying specific colours and co-ordinates on a canvas coming together to contribute a dot each and making Water Lilies (which too isn't as simple as it looks, but with some help from an ever growing community of Managers equipped with Excel, iPhones and Twitter, I don't think it should be a problem).
In one of the photographs of the Trevi fountain, one can also see a crowd that forms a pixel cloud/ghost that works like a contribution of remnant reverberation or rather spectral residues of different users who photographed themselves at various spots. Also, based on the popularity the city's visual modeling changes pixel density, and in some places dissolving completely in thin air.
No comments:
Post a Comment