You know the architecture of Paris when you see it. But what makes Parisian architecture distinctive? What visual elements come together to produce a particular urban landscape? Five scholars from Carnegie Mellon and the École normale supérieure are trying to give precise answers to those questions. And they’re taking a novel approach. They’re running a large repository of geotagged imagery from Google Street View through a proprietary algorithm and then identifying the distinctive architectural elements for each locale — the street signs, windows, balconies that make a city unique. Their experiment (all summed up in a short abstract here) covers Paris, London, Prague, Barcelona, Milan, New York, Boston, Philadelphia, San Francisco, San Paulo, Mexico City, and Tokyo.
Google “Art Project” Brings Great Paintings & Museums to You
Google Street View Opens Up a Look at Shackleton’s Antarctic
Tour the Amazon with Google Street View; No Passport Needed
Great. Very creative use of the GSV…
The picture on the left at the very beginning and end of this video looked so much like a place I sat one day, I had to hunt down my photos from 10 years ago. It’s not the same place, but oh-so-typical!
Very interesting use of Google photo records but note that the means of identifying cultural differences is by using examples from history. Try doing the same with modern architecture and you will find that it would be an impossible mission due to the fact that architects have stubbornly persisted with a ‘global’ or ‘international style’, thus denying any expression of culture.
nice.. but wait a minute…
That’s the same lab I contacted about more than 1 year ago about one of my ideas (applying visual computation technics on google street view imagery to create perceptual city indicators) applied on the same city (Paris). Strange coincidence…:(
here the link of my original work I finally published on my blog 1 year later: