Vodafone's Otello search engine uses images, not text

We've certainly seen some clever methods of searching from one's mobile, but Vodafone's latest idea is quite the stroke of genius. Showcased at CeBIT, the Otello search engine simply uses images as input; in other words, handset owners just snap a picture of anything -- a landmark, DVD case, unidentified flying object, etc. -- and Otello then "returns information relevant to the picture to the mobile phone." Reportedly, Vodafone is expected to conduct a trial with German paper Bild in which readers can "find out more about specially-marked articles by photographing them with their mobile's camera and sending the image to [the aforementioned paper]." Unfortunately, the carrier is being tight-lipped with its plans for Otello beyond the trial, but if this stuff functions as advertised, we can't see it remaining a secret for long.

[Via Pocket-lint]

[Grizz @ Mar 4th 2008 9:44AM]

Thats the same way I check my fly to see if its down.

[MikeG @ Mar 4th 2008 11:49AM]

Smart moves, there should be a website for guys full of this usefull info :D (Hint: If it exists, tell me)

[dervheid @ Mar 4th 2008 9:50AM]

"find out more about specially-marked articles by photographing them with their mobile's camera and sending the image to [the aforementioned paper]."
That's impressive, how?

[Carbonize @ Mar 4th 2008 9:52AM]

Sounds like a more complicated version of Qode.

[wxjunkie @ Mar 4th 2008 9:54AM]

Or like CueCat, heh.

[beaubrewer1977 @ Mar 4th 2008 9:54AM]

So, you take a pic of the paper and the paper returns the result. Sounds like a new twist of the old Radio Shack's CueCat. Sounds cool, but I'm sure it will flop. But I concede, it's a new idea, and it will have tried several times before any real useful and broad impact will be made by the common user.

[patsy @ Mar 4th 2008 10:40AM]

How is this the same as CueCat, other than on a very superficial level? The CueCat required extra hardware and software, required you to be sitting at your computer (MAJOR downer!), and didn't do much else besides taking you to a website. This doesn't take any extra hardware beyond what you already have, and while it's also pretty much "single purpose", the purpose is much more flexible and generic: searching. It can potentially let you find things that are hard to verbalize into a query.

Still, I think much more flexible and powerful use of phone cameras (other than accurate OCR) could be made with the widespread use of 2D bar codes, which can encapsulate much more useful information such as URLs, vCards, iCals, or just plain text to be pasted into another app. Check out http://www.i-nigma.com/personal for some possibilities. While their reader software is clunky and bloated, something like that built into regular phones or into Windows Mobile Search could be extremely useful.

[beaubrewer1977 @ Mar 4th 2008 12:36PM]

Patsy...

The only real connection I intended to make was the probability of failure of the service/technology.

You said: "CueCat required extra hardware and software"
What about people who do not have a cellphone, then would that not be considered "extra" hardware? Software, okay, probably not with the Photo search.

You said: "required you to be sitting at your computer (MAJOR downer!)"
Okay, you win, here.

You said: "didn't do much else besides taking you to a website."
Well wouldn't the service return a similar "page" just in mobile form, via text or pic message. And more than likely the text would contain a "read more" link for your mobile browser. I do not see the difference there, unless you are only looking for a truncated story or article. (Plus if you are taking a pic of the story, could you not just read it there, while it's in your sight?)

You said: "This doesn't take any extra hardware beyond what you already have,"
So I did not have the CueCat I had to go to Radio Shack and get one; so what if I did not have pic/text messaging. I would have most likely to pay extra for that, CueCat was entirely free. But I suppose on the flip side with CueCat I would require a PC and a interent connection. So I will stay neutral on this.

You said: "also pretty much "single purpose", the purpose is much more flexible and generic: searching."
Well if the CueCat was to prove popular would it not be about the same. I'm too lazy to type or search for product number X or article X. Or, on the flip side, there is a pic of John Doe Celeb, lets take a pic and search, and see what he/she's doing.

You said: "t can potentially let you find things that are hard to verbalize into a query."
Key word "potentially". I think it would be awesome to do that, but like I said it is more than likely bound to failure for the 1st few times. There are cell phone with awful cameras, just like a subpar OCR, you are going to get inaccurate results. I think that in time and proper support it will work, but for now, site specific service results are too narrow, and not to mention, now they have your cell phone number, where as cuecat I just gave them a dummy email. I would rather get spam to a dummy email and to my phone. I check my phone and I disregard my dummy email.

You said: "flexible and powerful use of phone cameras (other than accurate OCR) could be made with the widespread use of 2D bar codes,"
Well with this photo search technology, could you now revisit the cuecat, and take a pic of the bar code? But in the larger scene, how will the search know Celebrity X from John Doe with similar facial features, with a blurry, poorly lit photo. But the same can be said with taking a pic of CueCat bar code. But I would imagine the CueCat bar code would prove to be more accurate, via pic messages.

You said: "encapsulate much more useful information such as URLs, vCards, iCals,"
Both could depending on the implementation of the service. CueCat could have registered your mobile number and you could have had an option to send the info from the scanned bar code to mobile. -Again depends on the implementation. Not to mention CueCat was 10 years (estimate) earlier.

So I guess in short I still stand by my comparison, the Pic Message Search is very similar to CueCat.

[ericisshort @ Mar 4th 2008 3:48PM]

GAH! I cant believe I just read all that. Probably the largest and most redundant reply ever.

[beaubrewer1977 @ Mar 4th 2008 4:38PM]

Laughs... well I guess I got the most redundant award.

[ReggieXuk @ Mar 4th 2008 9:58AM]

*snaps picture of a blonde*

"christina aguilera returned every time"

[ijyt @ Mar 4th 2008 9:58AM]

I want something like this for google.

[Liam Billington @ Mar 4th 2008 10:04AM]

Vodafone isn't new to the game. Look at Microsoft Lincoln and that technology Bill Gates demoed at CES this year.

http://www.modaco.com/content/Windows-Mobile-News/252971/New-from-Microsoft-Lincoln-Mobile-search-using-pictures/

http://www.engadget.com/2008/01/06/bill-gates-live-at-his-12th-and-final-ces-keynote/

[Jason @ Mar 4th 2008 10:10AM]

It looks a little bit like MusicID... should be interesting to see how well it works in practice.

[mymaclife @ Mar 4th 2008 10:18AM]

If I take a picture of a pair of tits does it bring up Gates and Balmer top of the list?

[peshue @ Mar 4th 2008 10:46AM]

Are these specially marked articles something like thse QR codes (I think that's the name) that have been all over japan for a few years.

[ET @ Mar 4th 2008 10:47AM]

I will finally be able to name porn stars.

[Dan S. @ Mar 4th 2008 10:50AM]

Most Frequent Search Term:

(o)(o)

[MJB @ Mar 4th 2008 12:22PM]

I need something like this for music. How often do you hear a song and want to know what it is, but can't find it? Often, unless it has lyrics. (for me atleast)

[Sky @ Mar 4th 2008 12:48PM]

It exists (In the UK anyway, and is quite good, I use it all the time):
http://www.shazam.com/music/portal/template/pages/p/what_is_tagging.html

[MJB @ Mar 4th 2008 12:56PM]

:bows

[sopan @ Mar 4th 2008 12:59PM]

This has been available by a company called Mobot (www.mobot.com). Few of the Mobot guys are also behind Song ID...

[mike @ Mar 4th 2008 2:55PM]

This is EXACTLY what the mobile Web is all about. Minimal surfing, maximum results.

[MangaMan @ Mar 4th 2008 3:04PM]

I'm just waiting for the searchable people database - snap a picture of that girl standing by the bar, and retrieve a list of her past lovers, interests, current friends, income level, etc. by retrieving data from the various social networks.

Searchable Chicks FTW.

[Ian @ Mar 4th 2008 11:36PM]

the ULTIMATE lazy stalker tool... you dont even need to log onto myspace or anything!!!

[David Gavilan @ Mar 4th 2008 10:13PM]

This reminds me to my own research 5 years ago. Except that I also used the GPS location to restrict queries (in the case of sightseeing information, not for portraits). I guess I shouldn't have given up with the project...
http://www.img.cs.titech.ac.jp/~david/research.html