I just found this thing:
Google Image Swirl
I don't know what the heck it's for but it's sure fun to play with.
I just found this thing:
Google Image Swirl
I don't know what the heck it's for but it's sure fun to play with.
Code://try //{ if (a) do { f( b); } while(1); else do { f(!b); } while(1); //}
Very interesting indeed. Seems like a context-based search for images. Probably analyzing tags, image color, shape?
Seems to currently have a problem with the same images on different websites being repeated across the whole... circle thing.
Originally Posted by brewbuck:
Reimplementing a large system in another language to get a 25% performance boost is nonsense. It would be cheaper to just get a computer which is 25% faster.
Reminds me of the reverse image search engine.
Consider this post signed
If you consider that to be a problem. For me it's distinctly the purpose of the whole technology.
In many cases, I'll see and image and want to know more about it but the current context gives little description. In many cases, I've had to manually use Google Images and refine my search based on size and color to find the image in other sources and hopefully find relevant information. This does all the work for me... unfortunately a few of the searches I have pending for such technology are not available in the demo.
Sent from my iPadŽ
The Google Image Swirl page is just a big white screen. Does nothing.
"Owners of dogs will have noticed that, if you provide them with food and water and shelter and affection, they will think you are god. Whereas owners of cats are compelled to realize that, if you provide them with food and water and shelter and affection, they draw the conclusion that they are gods."
-Christopher Hitchens
There's also http://www.tineye.com/.
Is that were you got that ugly avatar?
Originally Posted by brewbuck:
Reimplementing a large system in another language to get a 25% performance boost is nonsense. It would be cheaper to just get a computer which is 25% faster.
I would guess that images are converted via some FFT to extract frequency-domain parameters for luminance channel and some similar thing for the color distribution for chroma. Such parameters would be independent of image size (within limits), independent of image orientation (reflections, rotation angle), and even some basic linear distortions like aspect ratio mismatches.
Using this internal representation, a sophisticated search engine could determine the degree of "similarness" between images.
I tried the search engines mentioned but none seem to do what I want.
Oh and somewhere there are face recognition search engines as well. Although I've only heard of them within sub-domains like perhaps flickr, or a local photograph organizing application. Are there any that can search the WWW for "Joe Blo" for example, given a sample photo?