New Computerized System Estimates Geographic Location Of Photos

New Computerized System Estimates Geographic Location Of Photos

www.sciencedaily.com

ScienceDaily (June 21, 2008) — Researchers at Carnegie Mellon University have devised the first computerized method that can analyze a single photograph and determine where in the world the image likely was taken. It's a feat made possible by searching through millions of GPS-tagged images in the Flickr online photo collection.

(visit the link for the full news article)

This is very interesting and of course with considerable security and military applications.

www.sciencedaily.com
(visit the link for the full news article)

This isn't quite the same technology but it has some similar features.

In the video linked below, you can see a soon to be released product called Photosynth. In the TED demonstration, Flickr was used as a source for the program to scrape all tagged images of a certain cathedral. An algorithm then associates the images into a navigable panoramic scene in which you can move around, view single images etc. In one part of the demo, the software even found pictures of a mountain climber and associated them into a chronological sequence. The way the photos link together yet are still detectable as being individual is an amazing concept.

They claim that one day all of the outdoor images on the internet could be linked into a seamless google earth type scenario.

Very cool stuff, I HIGHLY recommend this clip:
PhotoSynth / Seadragon: TED Demonstration
Here is another clip of a demonstration done for NASA: NASA Clip

Edit: A pre-beta demo of PhotoSynth is available here. It's not as advanced as the version they are running in the clip but still very cool to experience...


[edit on 21-6-2008 by SystemiK]

I really don't understand the math behind things like this... I have a hard enough time figuring out how many fingers and toes I have... but I am fasicinated by it and understand enough to get an idea of how it works.

[edit on 21-6-2008 by grover]

I'm with you man, I don't even see how applications like these are even possible but they seem to be working very well and this type of tech is only in it's infancy. I'm sure that later iterations will be absolutely mind boggling..