If you have an iPhone, you’ll probably want to check out SnapTell Explorer, a free application now available on the App Store. The premise is simple: take a photo of the cover of any CD, DVD, book, or video game, and the application will automatically identify the product and find ratings and pricing information online.
I was skeptical when I first saw the app – the iPhone has had difficulty with image processing for barcodes, and most image recognitions systems I’ve tried on other platforms have been iffy at best. But SnapTell just works. Every time.
The app correctly identified just about everything I threw at it: Xbox games, Pocketbook O’Reilly manuals, The Dalai Lama’s Little Book of Wisdom, Kurt Vonegut novels, and a number of more obscure books (yes, it worked on The Twinkies Cookbook). It even managed to ID a copy of Civilization 4, despite the fact that it was covered in obnoxious price tags and stickers. I actually tried to mess it up by taking photos in poor lighting and odd angles, but the app still stayed nearly flawless. No, it doesn’t have everything – I managed to stump it on a book about Danish Grammar – but it will do just fine for any trip to a retail store.
But while SnapTell seems to have the technology perfected, the app itself still needs a little work. Once you’ve located a product there is no rating or description offered – instead you’re directed to the appropriate links on stores like Amazon and Barnes and Noble (it would be nice if some basic rating information was pulled into the app). There’s also currently no way to quickly view a product’s price across multiple online stores, though this will be included in the next release which is expected in the next few weeks. The UI could also use some more polish – buttons are oddly placed, and the app doesn’t look nearly as slick as it should.
SnapTell works best on Wi-Fi and 3G, but also supports Edge (it takes around 10-15 seconds to upload the image on the slower network, versus a moment or two). The application will also be coming to the Android soon, and will feature both the image recognition seen on the iPhone version as well as barcode lookup (which is popular on Android but very difficult to pull off on the iPhone). The app was developed by SnapTell, a company that primarily focuses on image-recognition based marketing, and is making use of the company’s 5 million+ product database.
I’m a big app fan, and it’s been a little while since I’ve been impressed by a new one, but SnapTell is incredible! I tried it with a couple things around the condo and it’s 100% accurate. Nice find.
We have now put in comparison pricing info for books/CDs/DVDs/video-games into the app. Hopefully this will make the app even more useful.
Downloading this now! Thanks for the heads up, can’t wait to try it out.
A very innovative application, indeed. I tried it and it seems to work reliably. It makes it certainly worthwhile to have an IPhone.
They spent about $100,000 to buy an annual subscription from Muze.com. If you put “Muze something something” for their marketing promotion, Muze would give you a $10K discount. Muze which crawls the Internet and other sources to build their database of products. Amazon, ebay, lala.com, and the likes use muze.com database. Muze only has DVD, books, CD, and movies.
It’s pretty impressive. Even if you obscure the image by putting your hand in the picture, it’s still generally able to pick up the right product. It was even able to differentiate a DVD from a soundtrack!
Meh. I’m not much interested in anything but Danish Grammar books, so this probably isn’t for me.
Thanks for the generous praise!
Don’t own an iPhone? No problem. You can access the same information on virtually any camera phone, just snap a picture and send it via MMS to fun@snaptell.com or use the shortcode SNAPIT (762748) on ATT.
Did you know sending an MMS cost about $1.45 per message? If you send 20, it equals to the price of a cable bill.
what? mms (picture message) costs the same as a text message. all included in the texting plan i have on the phone.
I can take out my phone, take a picture of the product I want to buy, wait for feedback and price…or just pick it up, read the box and scan it at the register?
I don’t think the back of the box will give you objective consumer reviews, or competing stores’ prices, but I wish you well on that endeavor.
You are so stupid about objective consumer reviews. Big companies pay bloggers and people to write good about their product on the Internet.
cool app
kooaba (www.kooaba.com) provides a similar service.
Two thumbs up!
Not finding the app very good. I tried it on 5 different things and only had one success. I snapped a copy of jQuery in action, which it found. Everything else that failed included pics of a Sunkist can, a Coke Zero can, a copy of Revenue magazine, and a box of Quaker Oats from the company kitchen.
“take a photo of the cover of any CD, DVD, book, or video game, and the application will automatically identify the product and find ratings and pricing information online.”
I think it’s because it wasn’t made to identify things like that. It’d be like taking a picture of your sock.
The technology works on image differential. What it does is compare your image against the images in its database.
Snaptell is not smart enough to identify products beyond the specified 4 above.
So Mike, which competitor do you work for?
Downloaded it. Tried it. Wet my pants. This is VERY COOL!
Awesome app, This is one cool thing I have come across. Helps especially when I am looking out for movie details. Matches my even poor quality images.. amazing technology!
Wow, I just tried it. The first time it pulled an error because I think the server was too busy. But the 2nd time it worked perfectly.
That is one neat idea. I am getting some new ideas from this one.
impressed. i assume this also runs in ipod touch. won’t it?
a touch doesnt have a camera but it would work on photos already in the library
I don’t understand your comments about bar codes. What is it about the iPhone that makes it bad at handling bar codes, and what about Android makes it good? How is the bar code accuracy not just a function of the camera quality and the algorithm used to decode the image, which surely can’t be OS-dependent?
From what I’ve been told it has to do with the iPhone’s lense and the way it focuses. It is apparently possible to design an algorithm that works on the iPhone for barcodes, but is more difficult.
Wait are you saying that the iphone is unable to capture a clear picture of a barcode?
I mean after all its just about capturing the picture. Once the picture is captured successfully (without aberrations), the algorithms applied to it should not depend on what device was used to capture it… no?
jamboree: The iphone’s camera cannot focus on near images; the G1
jamboree: The iphone’s camera has a fixed focus; the G1 not only has adjustable focus, it can autofocus. Therein lies the difference; the iphone produces exceedingly blurred (read: loss of information) shots of barcodes while it is clear on the g1.
Would it recognize shoes, clothes?? And why not extend the service to instant group buying, aka “tuangou”.
I wonder why Like.com hasn’t come up with something like this yet?
There already is a like.com iPhone app. Check this out -
http://partner....urrentStep=SADS
@jamboree Like.com will probably buy this app!
personally I find this quite unnecessary… but whatever works
If only you could snap the food on another table and it would say which meal it is off the menu.
Or you could just develop social skills and politely strike up a conversation with a fellow diner…
Wicked cool. It recognized 100% of all books and DVDs I’ve hit it with. Forget Like.com …. Amazon should acquire this ASAP.
Oddly enough, it recognized my brother’s book that was published last week, but failed to match (a) Martin guitar strings, (b) staples, or (c) a can of Dust-off.
That isn’t odd. It only works for CDs, DVDs ,video games, and books…
This was my finishing project. It additionaly parses info directly (for example rating of a movie from imdb) and shows on app.
Reminds me of TinEye Mobile, which works for CD covers, and has direct links to the iTunes music store.
That must a very powerful image recognition technology behind it. It also means that scalability may be a issue as it requires massive processing for such technology to work. A lot of variables come into play like color, size, shape etc.
It just failed to identify a can of 7up.
For all of you who are taking pictures of random objects, it’s designed to only recognize photos of a “CD, DVD, book, or video game”. If it isn’t one of those, it’s not going to work.
Can’t find my wallet, pretty sux.
I tried several common items including Windex, and took several different unobscured photos in good light, from different angles and not one match.
CDs, books, etc.
Windex and common items — NOT part of the app.
For anyone who’s curious, check out this Boston based competitor: http://www.pongr.com
Works well enough on US products with the original US designs/artwork on them. It’s common however for the artwork to be changed in other countries though. In Canada for instance, french text is often added – and this causes a problem for the software.
Still, it’s promising…
My new Shopping buddy!
Sound like it still have a long ways to go. How would I scan in a lawn mower. Nice attempt though. Wait until my product based program comes out for the iPhone, it will blow this one away.
Ok I snapped just 3 different books, 1 book was a new york bestseller of last year.
And it didn’t recognize any of the books.
I love SnapTell! It’s saved me so much money! I actually just saw a pretty cool demo too, check it out! http://tinyurl.com/d32gjw