Occipital, a Techstars 2008 company based in Boulder, has released the first accurate UPC/EAN scanning application for the iPhone, called RedLaser. Users can snap a picture of a barcode and search for online product prices. They can also email a list of scanned products to review later or share with others. It’s available on the App Store for $1.99.
What’s interesting is the hard image processing problem they had to solve to make it work. The current iPhones lack autofocus, causing images of barcodes to be extremely blurry and tough even for humans to read.
In RedLaser, the blurry barcodes are sent to Occipital’s servers to be analyzed, and the system mathematically compensates for the out-of-focus images. Occipital will be able to improve the accuracy of the system over time as they get more data.
Occipital is also releasing an SDK for their scanning technology, allowing iPhone developers to integrate it into their applications. Cor.kz, the leading iPhone wine application, will be integrating RedLaser technology to allow their users to easily search over 600,000 wines.
This isn’t the first time they’ve pulled off some cool image processing. I previously posted when they showed off this demo at CEA’s iStage competition. iSmashPhone covers RedLaser with screenshots and more. RockyRadar has a good writeup with some additional information as well.
This can be really big, but I'm hoping Occipital changes their API scheme a little. I do a lot of business data work, usually on the order of moving data from one system to another. We get a lot of requests for barcoding. Often, they don't want to invest in dedicated barcode scanners, like those from Symbol. Either you buy a cheap scanner which uses a serial interface, and are tied down to a computer. Or, you buy one of the Windows Mobile/Windows CE based ones, and end up paying more than buying a laptop to tether a cheap one too. Much of the time, the barcoding is a minor part of the process, needed only once in a while, so the extra baggage is really unrealistic.
If we could offload the barcoding to existing mobile devices, it would save a lot of hassle. The customers already have cell phones with cameras, so there's no additional hardware to buy. Plus, they already carry their phones all the time, so you don't have to lug around extra equipment.
The problem is that iPhones are only a small part of the market. Blackberries are still more common, and you still see Palms, Windows Mobile and legacy phones. Restricting the market to just iPhones pretty much kills the usefulness of the app.
If instead/as well Occipital were to support uploading via an HTTP post, it would be a pretty easy thing to build simple web pages that any cell phone could use. It wouldn't be as slick. You'd have to take a picture, then go to the web browser and post it. But, that's not really a problem for the use cases I've seen.
Billing would have to change. Model it on the SMS gateways (clickatel for instance). $0.005/scan, purchased in blocks of 1000 scans or whatever.
The idea is to increase the usage of the app by increasing it's availability and ease of use for all developers. The iPhone is a great platform, but why limit yourself?
Oh, and you know you'll *never* sell the product or company to Motorola/Symbol without Windows Mobile support. Yeah, yeah, I hate developing for Windows Mobile, too. But, Motorola has the money and they need something like this, so you might as well target them as a potential buyer. Heck, they'd probably drop a couple hundred thousand just to keep you from competing with their bread-and-butter scanners.
As more and more smartphones get adjustable focus lenses it would seem at first blush like there will be less need for Occipital. However, optical (as opposed to laser) scanning still suffers from the need to capture the image in a single shot, which means the user must inconveniently frame the barcode, as opposed to just waving the product at the scanner, or vice versa, until it registers.
My hope is that Occipital (or somebody else) will address this problem by stitching together partial images from a live video feed to make the experience with a phone more like the experience with a dedicated laser scanner.
The point Dave Rudder makes is also valid, and suggests that the scanning (or other identification) of products should be separated from the retrieval of product information once the item is recognized. The applications we have now are analogous to AOL, in which the client application and the database were tightly bound.
I need the cor.kz wine-label iPhone picture capture catalog/price-compare/cellar-history/notes-history (all based on the label snap) app yesterday! bring it! happy to pay $30 on-time for the app, and/or a monthly ($10 or less) subscription. just give me the app.