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

What I need

[Intron]

... is a program that will remember the names of the people in the photos.

Re:Where pictures are taken

[elguillelmo]

then... if there are 6 sources of pictures, by blindfold guessing you'll get it right 16.66..% of the time

Blue screen your pictures

[BMonger]

I propose we all take pictures with blue screen in them (not the whole background, just "enough") and then write a script to randomly replace the blue screen with alternative locations every time the picture loads.

Re:Where pictures are taken

[elguillelmo]

I'm not with you in the argument. Assuming there are just 6 cities, and that the proportion from each is the same: 1/6, if you guess randomly you are right 1/6 of the time. It's just like a die... Then, if there are zillions of sources but only six cities amount for most of the pictures, then randomly guessing among them will get you close to this 1/6...

Automatic image recognition is no walk in the park

[hedu]

Reminds me of the experiment done in a Dutch military lab a couple of years ago. They trained a neural network to recognize whether a photograph taken out on a country road had a military vehicle in it or not.

The system recognized the photos from the training set perfectly, but did no better than random on images fed to it that were taken at different times.

Turns out all the training shots with a military vehicle in it had been taken on a sunny day, and the control shots without one had been taken when it was overcast. The system had been trained to recognize a different thing from what they intended!

Google

[sckeener]

Google should get behind this. I think their Picasa would benefit from it.

Generate some autotags.

What would be nice also is if they had a feature where if you labeled someone in a picture, if you uploaded another picture with that person in the picture, the program would prompt to auto tag.

I've been going through old family photos and it would save so much time if the programs I am using autolabeled based off details in the picture.

Re:Missing double blind

[Bat Country]

I'd imagine great work could be done by examining light intensity and coloration (atmospheric red shift) vs date stamp on the image (working from RAW with some camera data), they could guess the latitude fairly accurately. By similar methods you could figure out pollution levels, thus narrowing the sample range further.

Additionally comparing geometry could help factor out region with plant recognition fairly well also. You're not going to see a saguaro in Kentucky unless you're in a botanical garden. They've got a rather distinctive shape, and somewhat unique coloration.

Then you've got horizon lines - they're going to be ragged everywhere.

City skylines can be fairly easily identified the same way barcodes can be recognized, and mountain ridgelines are equally useful. The real trick would be telling a place in western Montana in mid-spring vs a place in western Kansas in early fall.