Google Goggles: Take Photos of Things Around You to Search for Them

Google launched Goggles, an Android phones application which lets you snap a photo somewhere, to then get more information about the contents of the photo. (“Reverse image search” engines like TinEye.com turn out to be very useful at certain times.) Some supported use cases for this, according to Google, are:

  • Snapping a pic of a landmark, like the Golden Gate Bridge, and then seeing the web results for that query “Golden Gate Bridge” (like a pointer to Wikipedia’s entry on the subject)
  • Photographing the cover of a book to then preview the book at Google Books, look for its prices at Google Product Search, or check the web results for this title (e.g. its Amazon page)
  • Taking a photo of the name and address printed on a business card, to then find the contact details be broken into their structural parts like email, phone number and so on, ready to be added to your address book
  • Taking a photo of an artwork to find out the name of its painter
  • Dynamically getting location places when you point the camera towards restaurants, cafes and bars on a street
  • Taking a picture of a wine label get more info about the wine
  • Taking a picture of a logo to find out more about the company behind it

I haven’t got an Android phone (yet*), but Mathias Schindler wrote in to tell me this app is “the best thing since sliced bread”. Pd comments that “not only is one feeding of the database, one is also adding to it.. Big advantage to Google!” Mathias adds:

I “goggled” a couple of Van Gogh paintings, all were recognized. In one case, instead of the painting a book with that cover was recognized, with the original painting as alternative. The ISBN bar code of a German book didn’t work, several Wikipedia pictures worked. It’s also very interesting that the quality of a snapshot of the computer screen was sufficient to return reasonable results. The business card scan worked very well.

Hebrew text recognition at the moment didn’t work at all, other non-Latin character sets I didn’t yet try. No doubt about it, the currently missing next step for Google would be to translate recognized texts, too. The technology is already available at Google after all.

If we continue that thought, then we end up with a user interface that could run even smoother than now. No more changing between camera and result, but, for instance, the embedding of the translation of a sign right into the camera picture, at the appropriate place. If Google’s product videos showing the augmented reality examples are to be believed – I didn’t try that part out yet – then this is already partially possible.

Is this a killer app, e.g. a piece of software that on its own can convince people to buy a specific phone/ OS/ hardware? Have you tried it, and if so, what are your thoughts?

[Thanks Pd, BizAbh, Mathias, Luka and Mbegin! Left-hand image CC-licensed by
Brewbooks, other images taken from Google’s product video.]

*It might be nice for those of us without an Android phone to have this as normal web app too. Is Google trying to get more Android phones sold, or are they afraid to scale this app for all web users, or don’t they think it’s useful in a desktop browser… or what could be the reason for a mobile app only release for now?

[By Philipp Lenssen | Origin: Google Goggles: Take Photos of Things Around ... | Comments]

[Advertisement] Search Marketing Arena: SEM/ SEO Questions & Answers. Increase your skills. Join now