Google Goggles

Maybe I’m late noticing this new feature, but I’ve just discovered Google Goggles. It’s a beta implementation of something I’ve wondered about for a long time: imagine a reverse-lookup engine for Google Images. Instead of typing in text to find an image of what you know you’re looking for, Google Goggles lets you take a picture of something and search for what it is. That is, Google Goggles is “visually searching.” Combining OCR (when there is text in the photo, like on a business card) with object recognition (e.g. bar codes), Google Goggles must cross reference your photos with a huge database of landmarks and icons. Demos on the site show it working on the Eiffel Tower and Transamerica Pyramid.

It currently works on both the Android (min. 1.6) and iPhone (iOS 4) platforms. Goggles is part of the Google app. So, if you have already installed the Google App, you may already have Goggles on your phone. If not, then update the app, and it should now be one of the search options.

After discovering it on my phone, I ran a few tests myself. Using what is available in my office, I wanted to see how well Goggles works on a few of my everyday objects.

MacBook

Not bad. Although my laptop is actually the 15in model, I’m not going to quibble.

real book

It recognized the first book I found handy.

notebook

And if searching a book with a title emblazoned across it seemed too easy, I next took a picture of my moleskine notebook. You can see here, it did not recognize it.

coffee mug

Searching my office for logos, this is the only one I could find. Sticky Fingers Bakery is a vegan bakery in Washington, DC. And although Google did not recognize the logo, it must have done OCR on the text.

As one of the developers says in the video (below), this technology is new and has a long way to go. Nevertheless, I see already some strong potential and many uses for this, especially as it gets better identifying art works.

How do you think you might use this new technology?

 

Leave a Reply