Register for an account


Enter your name and email address below.

Your email address is used to log in and will not be shared or sold. Read our privacy policy.


Website access code

Enter your access code into the form field below.

If you are a Zinio, Nook, Kindle, Apple, or Google Play subscriber, you can enter your website access code to gain subscriber access. Your website access code is located in the upper right corner of the Table of Contents page of your digital edition.


What Makes a Paris Street Parisian? This Program Knows

80beatsBy Sophie BushwickAugust 9, 2012 8:54 PM


Sign up for our email newsletter for the latest science news


The particular street signs, windows, and balconies on this street mark it as Parisian.

Paris, the city of light, is instantly recognizable---as long as you're looking at a photo of the Eiffel Tower or the Louvre. But could you recognize the city if the picture lacked a flashy landmark? (Try testing yourself here

---just don’t look at the text on signs.) If you find yourself stumped, know that a new software program has you beat: it can identify a city from a single photo

 of any old street. Almost any street, the program's designers found, has little details that give away its city, including distinctive street signs, windows, and balconies. To program the new software

, researchers at Carnegie Mellon fed it 40,000 Google Street View images from 12 cities, including Paris and New York. These photos contained hundreds of millions of unique visual elements, creating a huge database that allowed the software to pick out which recurring details were typical of which cities. For example, Paris has unique cast-iron balconies that are clearly distinct from New York’s fire escapes and London’s neoclassical columns. Mining large databases to find patterns like these is not a new technique, but the databases usually contain text or numbers, not pictures. Visual data mining is a much newer and trickier method, which might someday be applied at smaller scales to identify individual neighborhoods from a photo, or at larger ones to find the representative details of an entire continent. It's interesting to wonder whether, in the future, machines will pick up on distinctions that even a city's inhabitants are not conscious of, unique details that they walk past every day but never notice. Images courtesy of Carnegie Mellon

    3 Free Articles Left

    Want it all? Get unlimited access when you subscribe.


    Already a subscriber? Register or Log In

    Want unlimited access?

    Subscribe today and save 70%


    Already a subscriber? Register or Log In