By Andrew Liszewski
Those self-serve check-outs that have been popping up at more and more supermarkets are definitely convenient, but there are still a few technological hurdles to overcome that will make them a lot more efficient. At the moment, items like fruits and vegetables that don’t have a barcode and need to be weighed tend to slow down the checkout process, since you need to specify what they are via a touch-screen menu interface. (As opposed to just scanning the item and throwing it in a bag.) But thanks to a recent development from the Fraunhofer Institute for Information and Data Processing and Mettler Toledo, new check-out scales will be able to automatically distinguish between an apple or an orange. (Or whatever fruits and vegetables the store happens to sell.)
The scales use a digital camera to photograph the item, and an algorithm is then used to compare the image to a database of previously photographed fruits and vegetables to extrapolate what it is. In the case of something like tomatoes, which come in different varieties but might all look the same, the customer is simply presented with a set of different icons on the screen and is asked to specify between hot house or field varieties. The system is apparently even savvy enough to recognize items that have been placed in a semi-transparent plastic bag, and has a high tolerance for color and brightness fluctuations resulting from varying lighting conditions. The new scales are currently being tested in about 300 different supermarkets across Europe, and will no doubt eventually make their way to North America.