If you know or come into personal contact with an Android user, you’ve probably had that nifty little Google Goggles app thrown in your face a few times already. It’s a handy little app that uses visual recognition to identify objects and products and direct you to the Internet to find and buy said objects. Think of it as one of those barcode scanner apps, minus the need for a barcode.
Well, the day has finally come for the iPhone to get a visual search app of its own. In this installment of “Does it Work?” we take a look at oMoby, which hit the App Store a few weeks ago and claims to make searching simple by letting you snap photos of any product and locating it for purchase on the Internet.
Think oMoby will work? Check out our assessment after the jump.
How it Works:
According to the developers’, oMoby works a little by computer vision and a little by “magic.” When you snap a photo of an object in the app, the “computer vision” goes to work to identify the object, and when that fails, the app turns to a network of humans to help solve the mystery. The instructional video posted on YouTube is a bit more explanatory, as it describes how the app looks for characteristics in an image, instead of mainly text or logos.
When you open oMoby, the app launches right into camera mode, ready for you to snap away. I started with a few simple media items: a digital camera, a DVD (Teenage Mutant Ninja Turtles), and a paperback book (Into the Wild) to test the waters without rocking the boat.
Immediately after taking a photo, the app reverted back to my “history” screen, which displayed all of the items in my list of queried items. For each item, a thumbnail shows to the left and text to the right would read “Searching” and “Analyzing Image.”
After a few minutes, the text changed to read “We will notify you when the product is identified.”
After another few minutes and after closing the app, I received a push notification on the icon on my home screen, letting me know that an item I had searched for had been identified. The digital camera I took a photo of had been identified..correctly!
Opening the results page for the item shows a list of options, such as sharing your result via e-mail, or searching information on the item through sites like Google, Bing, and Wikipedia. The real meat of the app comes in the list of shopping links offered for 10 websites like Best Buy, eBay, Walmart, Google Products, or Overstock. You also have the option of posting your results on social sites like Facebook or Twitter. Even better, all of this searching on other sites is done inside the app.
So oMoby successfully identified my digital camera, but the rest of the media items didn’t go over so well—at first. It took much more time than the digital camera to identify at all—by about half an hour—and when the app finally did identify them, they showed as incredibly vague titles: “book,” “comic book” (for the DVD—wrong).
So I did more searches of some more difficult items, using the tips offered in the instructional video, and also retook photos of the items that were vaguely identified. The same vague identifications were given for these and took almost twice as long to get:
Lysol cleaning wipes — “Food Item”
Burt’s Bees hand lotion — “Bees”
Picture frame from Target — “Wood Thing”
Sharpie marker — “Sketch Pen”
Apple iPhone earbuds — “Earphones”
I’m not kidding about “Wood Thing,” either.
However, literally an hour later, while I was putting together this here review you’re reading, I checked my iPhone to see that I had 10 push notifications from oMoby. I opened the app to see that every item in my history had been updated and that each item had been more specifically identified, with titles for the media items and more specifics on the others: “Lysol Dual Action,” “Burt’s Bees,” “Picture Frame,” “Sharpie” and “Apple Earphones.”
I guess it really does work by magic.
In short, yes, the app does work. Patience and precision is absolutely necessary for the app to work as promised, though, and seeing as this is a fairly new app, it’s going to take a bit to work out some kinks. oMoby insists that the more the app is used and the more users correct misidentified items, the better the app will work. So we have to kudos for the “community” aspect to the app.
As a tip, do take the developers’ advice to take as clear of a photo as possible, to make things easier for the magical elves working behind the scenes. And although oMoby insists text isn’t necessary to make the app better identify objects, if you’re snapping a photo of a product with any branding or specific identifiers on it (i.e. the picture frame from above), you’re most likely not going to get the results you really want. Sure, if you snap a photo of a picture frame, you should expect to have the app identify it as a picture frame. But that’s not quite what users of this app are using it for, anyhow. The magic seems a bit lost, in this sense.
The final verdict? … It works. Go download it!