Back in late 2010, Google brought Google Goggles (not to be confused with Google Glass) to the Google Search app for iPhone. At the time, it seemed like a great feature. You could take a photo of anything with your iPhone, and Google would instantly identify the object in the frame, returning relevant data about the photo.
That was 2010. You’d probably think that in the years following, Google would’ve continually updated Goggles, making the recognition more accurate and expanding its features. Well, you’d be wrong. In fact, Goggles functions entirely the same as it did in 2010, another fine example of Google’s continued dedication to their products.
Image Searcher, the developers behind a new app called CamFind, have seemingly stepped in where Google failed, creating a image-based search service that they claim is “four times more accurate than Google Goggles.” Let’s take a look at how it really performed.
To see CamFind in action, check out this video overview of the app and its features:
CamFind claims to be able to identify a plethora of different objects and images, from movie posters to QR codes. Upon launching the app, you wouldn’t guess that it’s very powerful, but once you start searching, the real power of CamFind’s image recognition starts to show. CamFind’s interface is incredibly simple. The app looks almost like a run-of-the-mill camera app, until you start poking around. The bottom of the app holds just three buttons. From left to right you’ll find a history button which calls up a list of your past searches, the shutter button, and a menu button which triggers a small list of options that allow you to turn off the app’s built-in speech feedback, start a voice search, change the language for translation purposes, and bring up the about screen.
Taking a photo is no different than snapping a quick shot in Camera.app. What is different, however, is what happens afterwards. Immediately after taking the photo, CamFind will upload your image and start to identify it. This is where I began to run into issues. Although the app was typically able to provide feedback within a matter of seconds, sometimes identifying the photo would take a considerable amount of time, even on a fast Wi-Fi network. Occasionally, the identification process would time out altogether, and I’d have to start it again. This is the one and only place where Google Goggles takes the edge over CamFind. Even though the results that Goggles brings back are often times extremely vague or inaccurate, Google has found a way to deliver those poor results almost instantly.
After CamFind has identified your photo, you have a few options. Swiping across the small bar that overlays the top of the screen reveals the ability to delete the image, rename your search, share your search, or have the app speak the result again. Tapping on the result itself reveals an entirely new view, which offers up a price comparison of the item in question, related businesses near your location based on Yelp reviews, an image gallery of related photos, and various links that match the text description of your photo.
CamFind (left) and Google Goggles (right) displaying results for the same image
Upon initially testing CamFind, I figured that its accuracy was pretty close to that of Google Goggles. Then I remembered that I hadn’t used Goggles in a while, because it was so bad. So I fired up the Google Search app, and tried taking photos of the same object with CamFind and then with Goggles. It turns out that CamFind shames Google pretty badly. Even when Goggles was able to identify the objects I tested, it often did so in a vague, broken way, providing barely enough information for me to find out more about the product, especially if I was trying to identify something I’d never seen before. CamFind, on the other hand, was able to consistently provide detailed results.
I was particularly impressed by how descriptive the app was. Taking a photo of an old Mac of mine managed to yield the result “1980s Macintosh computer,” and snapping a photo of a painting of Apple’s headquarters returned “painting of trees and beige building.” Although I wasn’t able to get such accurate results every time I took photos, I was still thoroughly impressed. The one feature I wasn’t able to get working was barcode identification. CamFind would typically just return with the generic answer of “barcode,” instead of the product matching the barcode itself. Whether this is a flaw with the app, or something I’m doing wrong isn’t clear to me.
I headed into reviewing CamFind with some negative experiences already biasing my opinion of camera-powered searches. Although I found Google Goggles to be a neat tool when it first launched, its lack of accuracy quickly caused the magic to wear off. CamFind actually makes searching with photos fun and useful. I consider myself an avid complainer, and aside from some speed issues, I can’t find anything negative to say about this app. CamFind is free in the App Store, and I can’t recommend it more highly.