Computer Scientists Scour Your Holiday Photos

Barence writes "Hundreds of thousands of images on Flickr are being used to teach a program to determine the geographic location of an image, simply by looking at it. The program attempts to mimic the way that humans can deduce the location of an image by searching for visual clues, such as similarities to pictures or locations they have seen previously. In its current state it can guess the location of a photo to within 200km, 16% of the time — extremely accurate given the complexity of the problem."

Comments

Where pictures are taken

[tomalpha]

The paper referenced in the article has an interesting density map of where their 20 million source photos were taken (ok, so they only ended up using 200 or so of these). It says it uses a logarithmic scale, and seems to imply that the vast majority of photos available to them on Flickr were taken in one of only a handful of locations:

• London
• Paris
• New York
• Washington
• Los Angeles
• Tokyo

Ok so there are a couple more than this, and my geography is appalling, but these seem to be the only areas that are are coloured red.

Re:Where pictures are taken

[Ethan Allison]

There are way more than 6 sources, check the map.

Re:Where pictures are taken

[hagnat]

yeah, but only if the program chose the same city 100$ of the times, which i guess the guy who created this program would realize its malfunctioning and then fix it.

Assuming that, from the 200 pictures set used to validate the program, the same amount of pictures were provided for each city (32 pics of each city) and that the program chosen then in the same proportion (the program identified [correctly or not] each city 32 times), you would have 16.6% * 16.6% = 2.75% chance of correctly identifying the city where a picture was taken

Re:Where pictures are taken

[Falkkin]

If you RTFP, Figure 6 shows the difference between the performance of the algorithm and random guessing. It's pretty significant.

Photosynth looks cooler

[Bombula]

The Photosynth multi-resolution and image-recognition tech demonstrated at TED looked cooler if you ask me:

metacafe link here and TED link here.

Re:This is very hard

[Anonymous Coward]

I've already modded you so posting anonymously to preserve the mod.

To answer your question, yes, I would but only because I would know what to look for. In your case, the walls are not steep enough, too much vegetation and no thin grey haze hanging over the canyon.

Regardless, your point is still valid.

Re:Random pick is correct ~8% of the time

[Anonymous Coward]

Re-check your math that is wrong.
Should be .08%

Re:Lies, Damned Lies, And...

[SQLGuru]

Surface area of a sphere = 4*pi*r^2
Radius of the Earth = 6 378.1 kilometers (from Google: http://www.google.com/search?hl=en&rlz=1T4ADBS_en__230US231&q=radius+of+earth )

Surface area of Earth: 510,065,600 km2 (http://www.google.com/search?hl=en&rlz=1T4ADBS_en__230US231&q=surface+area+of+earth)

Percentage of surface area that is land: 29.2% (http://www.physicalgeography.net/fundamentals/8o.html)
Surface area of Earth that is land: 148,940,000 km2 (same source)

Area of a circle = pi*r^2
Radius of "target" = 200km
Area of target = 125663.7km2

Number of "target" areas that could fit on the surface of the Earth covered by land (assuming too few landmarks to identify pictures take over water, so they will be excluded): 1185.2

Chance of being right by pure dumb luck - 1 in 1185.2

Layne