Half a croissant, on a plate, with a sign in front of it saying '50c'
h a l f b a k e r y
Left for Bread

idea: add, search, annotate, link, view, overview, recent, by name, random

meta: news, help, about, links, report a problem

account: browse anonymously, or get an account and write.



Reverse Google Image Search

The quick and lazy way to identify an object.
  (+13, -2)(+13, -2)
(+13, -2)
  [vote for,

So, basically, this is google image search in reverse. You upload your image, and google tells you what it is, and maybe provide some links. This would work on the principle that robots can already identify objects, so the google servers could match it to a database of maybe five images per thing. So, let's say you find a cool bird in your yard, but you look in a bird book, and you just can't find it. Then, take a picture and use google reverse image search. Specify that it is a bird, and they will tell you what species and give some web pages.

by the way, I'm new here so tell me if this has been done.

dehodson, Jun 19 2008

Tin Eye http://ideeinc.com/products/tineye/
It finds variations of the image you upload. [Amos Kito, Jun 19 2008]

Not only what, but where also http://graphics.cs....du/projects/im2gps/
[AbsintheWithoutLeave, Jun 19 2008]

ahem... cartoogle
[po, Jun 19 2008]

Content-based image retrieval http://en.wikipedia.org/wiki/CBIR
[goodmars, Jan 14 2009]

RevIMG http://www.revimg.net
reverse image search with online paint, upload image or paste an image URL [Mauren, Jul 06 2009]

TinEye.com http://www.tineye.com
Reverse Image Search (Another link to it... because I missed the one by [Mr Kito]) [Dub, Jul 06 2009]

GaZoPa http://www.gazopa.com/
Another reverse image search engine. [jutta, Jan 03 2011]


       Tin eye is really limiting themselves by not taking pornographic images.
ldischler, Jun 19 2008

       +, but this is still a relatively new technology and I don't think that something as big as google could use it yet, it takes a while to teach the pc all the dif images. In 5-10 yrs, sure!
danman, Jul 06 2009

       Not so sure about that. 'Visions of the Future' showed ultra-capable object recognition machines that can beat a human, although I presume the algorithms may be relatively specialised for a particular task, to make it look as good as possible   

       (This one could recognise if there was 'an animal' in a picture, which is pretty generalised and impressive)
kindachewy, Jul 06 2009

       //although I presume the algorithms may be relatively specialised for a particular task, to make it look as good as possible//   

       Not really, they 'showed' the comp thousands of pics of named animals and then showed it entirely new pics of the same animals. The pc was then able to say weather there was an animal or not, faster than Michio Kaku!   

       I would like to alter my previous objection from genus species to a similar looking bird.
danman, Jul 06 2009


back: main index

business  computer  culture  fashion  food  halfbakery  home  other  product  public  science  sport  vehicle