A Japanese supermarket scanner is able to recognize subtle differences in produce without barcodes or manual data entry. Items are 'scanned' by being placed in front of a camera that can recognize patterns from a database, letting it ignore the clerk's hands or anything else that's not for sale. Its focus on color and patterning means that it's able to distinguish between extremely similar-looking produce varieties like Fuji and Jonagold apples. It's also able to recognize text, meaning it could be used to read coupons or scan packaged food. The scanner's creators are currently working on a comprehensive fruit database and fine-tuning the scanner so it can read items more easily.
In the video below, the system seems to work slowly compared to standard barcode scanners, and it may not even be faster than an experienced clerk who's memorized produce codes. But assuming it's able to cover every item in a store, it could speed up the checkout process for part-time employees or self-check systems, which usually either ask consumers to pick from a list of produce or manually input a code affixed to each fruit or vegetable. That means you'll have no excuse to opt for those easily-scanned chips next time you go shopping.