Finally Google Discovered Something and now it can read text from images !!

Update:
Please Digg this thread if you think it is really something new: http://digg.com/tech_news/Now_Google_can_read_text_from_images




I noticed this few days back and decided of sharing this with all of you !!

Yes this is true Check Google Catalogs
Search something ... uuum .. like i searched for "Computers"

and got this below Image Result with "Computers" Highlighted !!



Do can we conclude from this that Google Really discovered something with the help of which it can read text from Images ??

Read the full article here !

Let me know what you have to say about this

 

I wonder how that algorithm of the program works.
But it's amazing.

 

Pretty sweet that they have stuff which can read text off of images.

 

Google books have the same thing don't they? Open a book and use their tool to search it, you'll be impressed

 

is any s/w to read image text?

 

only Google can tell us about this !!

 

#1, it is visually unappealing.... as a long time C++ developer who has worked with image formats, music and video there is no way they can accurately accomplish it from only an image because they would have to be able to detect a font or font family then convert it to text unless the images have description information stuffed into them at the end of the file pointer.

Possibly, they are using GD or ImageMajik to compose an image on the fly from provided text/images and layout and then highlight the search word. Obviously it's still "alpha", not even "beta" because they still haven't figured out how to dynamicly write a corresponding imagemap with links to the items. Basicly, it's worthless unless a user can click on an item and get detailed info about. Seems pointless to me when it can all be done in html/php. Plus, using html, people with eyesight disabilities can resize their text. In an image like this, there is no text resize option....

 

Ever heard of Abby FineReader?

http://finereader.abbyy.com/?param=137503

 

why not?
babylon can read it...

 

I reckon it is pretty much similar. PDF's are read all the time and always seem to show in google's search results.

 

yeah that google books is a real pain in the arse, its always 2nd or 3rd in the serps and its never what im looking 4.

 

this ability has been available in many software before Google did this. It's not impossible at all, in fact many programs that come with scanners have this feature.

 

The technology is similar to those which use captcha recognition. Matt Cutts talked about this earlier and how they got a good laugh of an email which said something like use this spamming technique, its a picture and google can't read words in picture...

 

What so ever it is this is really a step ahed from google !!!

 

I don't think they can yet.. I think the game, image labeler that they have offered people to play, someone wrote that tag, and so it came up there..

 

LOL what about CAPTCHA reading softwares?

 

Try to search for "four computers" then.

Wow, amazing. And scary

 

Lol.. thats pretty damn cool. I am sure they implemented this a while back and tried to get those webmasters who cheats with alt text and all. Well done again Google!

 

You're sure those pages aren't uploaded in pdf format? Google has been able to index and read pdf files for some time already.

 

No they didn't......as much it appeals to your fantasy about google the fact is they can't read images !

Check out the page and scroll down to "Catalog content provided by the merchant. Please verify pricing and availability before purchase."

After all it isn't so high tech ... is it ?

The highlighted words are probably created jpg output from pdf files. Technology that already exits.