Google Soon to Recognize Text in Images?

Friday, January 4th, 2008 by Roderick Ioerger

Google Soon to Recognize Text in Images?

Information Week has reported that in June of 07 Google filed a patent application, which has just become available, outlining a “method of optical character recognition in digital images.” The application seems to cover both static images as well as video. The ability to do this could radically change a number of existing Google services as well as again change the way the Internet marketing world interacts with images and video.

Being able to identify text in images and video would be an extremely large leap in search engine indexing technology. Being able to index the content of videos and other new social media would most likely either force a change in the way search engines have to build their SERPs or provide opportunities to build highly relevant image and video search functions that could generate significant relevancy from the text content within the media.

Clearly this would not be a complete solution for new service development, as many images and video do not contain text to index. Still as Duncan Riley at TechCrunch summed up nicely what I was thinking when I read Information Weeks coverage of this patent:

This will make every book in the Google Books database really searchable, with the next step being YouTube, Flickr (or Picasa Web) and more. The search capabilities of the future just became seriously advanced.

Social Media Monitoring in Just 60-Seconds. Guaranteed!

Similar Stories in: Legal, Research | Forward: Email This Post

Share this post

35 comments on “Google Soon to Recognize Text in Images?”

Alan Johnson Says:

January 4th, 2008 at 11:22 pm

Recognizing text in images will definitely be revolutionary and create quite some buzz, if they manage to pull it through this year, it will definitely be a strong candidate for 2008’s most important Internet event.

Alan Johnson
Edward Says:

January 5th, 2008 at 1:23 am

If they managed to do that then it will be easier for users to get more targeted images, nice job Google!
Matt Fritz Says:

January 5th, 2008 at 9:43 am

I wonder if that means soon spammers will have something like that to read and bypass the little CAPTCHA images
» Pandia Weekend Wrap-up Jan 5 Says:

January 5th, 2008 at 10:20 am

[...] Google Soon to Recognize Text in Images?Google patent for a “method of optical character recognition in digital images.” (Marketing Pilgrim Jan 4 2008) [...]
» Google’s New Image Search Patent Big News for Website Search Engine Marketing Website Marketing Journal Says:

January 5th, 2008 at 11:46 am

[...] an article on Marketing Pilgrim today, Roderick Ioerger discusses a patent that was registered by Google approximately six months [...]
Famous Quotes Says:

January 5th, 2008 at 11:52 am

Google is always trying to do something new and they will most likely accomplish it. Like Helen Keller said, “While they were saying among themselves it cannot be done, it was done.”
Ben Cook Says:

January 5th, 2008 at 1:26 pm

Matt, that’s exactly what I was thinking about. Would seem like it would be the next logical step.

Also, I wonder how this would play out. I mean would it replace the alt text in the SE’s mind? Or is this just another step in trying to be able to identify images?
Alan Johnson Says:

January 5th, 2008 at 2:08 pm

Who knows, if people interested in bypassing the captcha image were to have access to this technology and use it to their own advantage, audio verification would most likely be an option.

Alan Johnson
Roderick Says:

January 5th, 2008 at 3:57 pm

Ben,

I was wondering the same thing about alt text when I wrote this article but the obvious answer and one I tried to include in my articles was that not all images or video have text to be read, so this would not be the death of alt text.

Roderick
Simon Heseltine Says:

January 5th, 2008 at 11:42 pm

This would be huge for online Yellow Pages, the ability to rip text out of the ads… this would give Google one huge leg up in local.
Suresh Chowhan Says:

January 6th, 2008 at 12:04 am

By this technique, search a relevant image will be easier for users. And make pages containing images spiderable.
Small Business Marketing Says:

January 6th, 2008 at 3:32 am

This is awesome technology and will most likely chew up cycles and bandwidth but who cares…..Now we can search a lot of the stored information that wasn’t cataloged with the Internet in mind.
Google erkennt jetzt Text in Bildern » SEO- und Internet-News by TechDivision Says:

January 6th, 2008 at 4:04 am

[...] ich gerade auf Marketingpilgrim lese, ist mittlerweile ein vor etlichen Monaten eingereichtes Patent verfügbar, dessen Technologie [...]
Ben Cook Says:

January 6th, 2008 at 4:29 am

But really, how often would you search for images by the text on the image? I mean sure some bloggers use images for their name of the blog or some of something along those lines but other than that, I really don’t see it being all that useful.
GoWFB Says:

January 6th, 2008 at 7:08 am

Wow! That would be amazing if they could make that possible!
Tom Melde Says:

January 6th, 2008 at 9:37 am

google does not cease to surprise!
Brian Chappell Says:

January 6th, 2008 at 11:17 am

The ability to crawl and read text inside videos is a big deal IMO. Brett Tabke pointed out this was coming about 7 months ago. Helps to know the right people doesn’t it
Wooden Pete Says:

January 6th, 2008 at 11:19 am

It would still leave it open to abuse, because although the software may able to recognize the text in an image, that doesn’t stop people putting text into an image that isn’t relevant to the image. Having said that, it is still a big step forward in indexing images.
Tom Melde Says:

January 6th, 2008 at 1:56 pm

google tries to fill immense and it at him not bad it turns out
Alan Johnson Says:

January 6th, 2008 at 2:05 pm

Ben, you know what they say: “a picture is worth a thousand words”, so that, sure, an image can be just as relevant as a 1000-word article

Alan Johnson
Nascar Says:

January 6th, 2008 at 4:26 pm

This is a great step forward by Google search. When the patented application can read Magic Eye 3D illusions, I will be really impressed.
Google to recognise text in imagery - more to follow? Says:

January 6th, 2008 at 5:46 pm

[...] post on the Marketing Pilgrim website quoted a recent patent application by Google outlining a “method of optical character [...]
Matt Fritz Says:

January 6th, 2008 at 6:46 pm

Also with Google wanting to scan all the worlds books, this would make it easier for them to search text inside a scanned image of a book.
Yiwu Says:

January 6th, 2008 at 8:30 pm

Oh,Its good news…Which I had looking forward to
Edward Says:

January 6th, 2008 at 10:02 pm

Great move by Google, it would be a great chance for photoblog to get more traffic
seonotes Says:

January 7th, 2008 at 1:24 am

It’s a good thing that Google is working on finding ways to improve relevancy of search.
Steven Bradley Says:

January 7th, 2008 at 2:00 am

This would definitely be a huge advance. So far what I’ve seen of OCR technology hasn’t left me overly impressed, though I admit it’s been awhile since I used any.

In regards to alt text I think we need to ask how much will Google’s technology cost? Will this be something that can easily and inexpensively added to screen readers? If not then alt text isn’t going away any time soon.

As for spammers and captcha it’s just part of the same game that’s been going on for a long time. Spammers figure out how to get around some technology so someone comes up with a new way to stop them. Spammers respond in kind. It’s the endless cat and mouse game.
Jim Hunter Says:

January 7th, 2008 at 4:48 am

This is not anything new. For years I have been able to sick OmniPage on an image and it can locate the text in it and OCR it out into a text file. And the latest OmniPage 15 (yes, it has been around for a LONG time) does an excellent job! So if this is something you have been waiting for, wait no longer, go get OmniPage 15. Google should not be awarded the patent unless they are using some new way to locate the text in the image, since prior art does apply here.
Zen Says:

January 7th, 2008 at 5:12 am

There have been some sites that were able to do image recongnition. Can’t remember any of them now… Anyway, wouldn’t Google Image Labler for texted images be cheaper? They could always give away a free lava lamp or something cool as a prize.
Mike Monaco Says:

January 7th, 2008 at 9:00 am

Captcha is already broken…http://sam.zoy.org/pwntcha/

OCR is nothing new.
Payday Land Says:

January 7th, 2008 at 11:08 am

That’s pretty interesting. I hope to see the progress in the months to come.
Alan Johnson Says:

January 7th, 2008 at 2:44 pm

Steven, I agree, spammers will always try to bypass everything website owners throw at them and, in the end, it is important to always try to be one step ahead.

Alan Johnson
Google Soon to Recognize Text in Images? - Google Community Says:

January 16th, 2008 at 10:37 pm

[...] Soon to Recognize Text in Images? guys have heard/read this… Google Soon to Recognize Text in Images? wonderin whats gonna be the effect of this? __________________ affordable web site [...]
What Does The Future Of Web Design Hold? Discuss | Web Design Blog Says:

August 20th, 2008 at 10:58 am

[...] January this year google also filed a patent for recognizing text in images, this would surely lead to more usage of images and changing the ways in which we edit and create [...]
jacky Says:

November 27th, 2008 at 1:59 am

This is not anything new. For years I have been able to sick OmniPage on an image and it can locate the text in it and OCR it out into a text file. And the latest OmniPage 15 (yes, it has been around for a LONG time) does an excellent job! So if this is something you have been waiting for, wait no longer, go get OmniPage 15. Google should not be awarded the patent unless they are using some new way to locate the text in the image, since prior art does apply here.