12

Does any one have an idea regarding what sort of algorithm might Google be using to find similar images ?

Emil
  • 13,159
  • 18
  • 67
  • 107

4 Answers4

4

No, but they could be using SIFT.

mpen
  • 256,080
  • 255
  • 805
  • 1,172
2

I'm not sure this has much to do with image processing. When I ask for "similar images" of the Eiffel tower, I get a bunch of photos of Paris Hilton, and street maps from Paris. Curiously, all of these images have the word "Paris" in the file name.

Niki
  • 15,476
  • 5
  • 45
  • 73
1

Currently the Google Image Search provides these filtering options:

  • Image size
  • Face detection
  • Continuous-tone ("Photo") vs. Smooth shading ("Clipart") vs. bitonal("Line drawing")
  • Color histogram

These options can be seen in its Image Search Result page.

rwong
  • 5,936
  • 1
  • 21
  • 46
  • No i talking about similar images. http://similar-images.googlelabs.com/ – Emil Aug 10 '10 at 05:22
  • 1
    I tried that tool and it didn't seem interesting at all. At best they're simply doing clustering using both word-image information (for example, linking the word "Paris" with all images appearing on web pages that include the word "Paris") and the image information (image size, face, histogram etc) – rwong Aug 11 '10 at 06:12
  • They may also take clues from http://images.google.com/imagelabeler/ – rwong Aug 11 '10 at 06:13
1

I don't know about faces, but see at least:

I have heard, that one should use this when comparing images (I mean: make the prob model, calc. the probs, use this):

Or then it might even be one of those PCFG things that MIT people tend to use with robotics stuff. One I read used some sort of PCFG model made of basic shapes (that you can rotate magically) and searched the best match with

Community
  • 1
  • 1
kummahiih
  • 56
  • 3