Do you ever think about how you recognize things you see? I suppose not, as it is such a basic human quality which we started learning from the day we were born. As a baby, we looked at toys and other—often colorful—objects, and we could look for hours at the same object, trying to absorb every tiny detail in shape and color. We also used our hands to feel the weight and shape, and matched that with what we saw. This is how we discovered the world around us, and learned to recognize all the things we now take for granted when we walk outside or watch a movie—or even when we read this article.
A thousand varieties of apples…
This is also what we do when we are in a grocery store. We look at our shopping list, and then quickly scan the shelves for visual clues for this specific item. Suppose we’re looking for oranges at the fresh produce area. By having a mental reference point for color, size and shape, we are able to distinguish between mandarins, mineolas and oranges. For apples, it may be a bit harder, as there are over a thousand different varieties. Thankfully, most stores don’t offer all thousand but (speaking for myself) even if you have to choose between 20 different varieties it gives you a hard time. Even for experienced store staff, it is often hard to find the right variety in the POS system.
Item recognition & self-checkout
Item recognition, in particular for fresh produce like vegetables and fruits, is often seen as a complicating factor for self-checkout solutions. We have to select the correct type of fruit and variety and then weigh it, before we can add it to the transaction. It’s a complex and time-consuming task that leads to frustration with shoppers, longer queues and mistakes on the receipt.
Smart vision technology for recognizing fresh produce
Fortunately, nowadays we have
advanced technology that can solve this problem: it enables customers to purchase non-barcoded fresh fruit and vegetables more easily. It can be automated using 3D cameras and deep-learning algorithms based on AI. It sounds complex, but the result is simple for customers. They just put their fresh produce item on a scale-with-camera. Smart vision technology recognizes the item and presents the customer with the proper name of the item. After customer approval, the fresh produce item is added to the transaction in a single click. This simplifies the checkout process for the customer while reducing staff interventions to help customers with incorrectly labeled fresh items. This approach to item recognition also contributes to improved stock accuracy. Using smart vision technology, consumer frustration can be replaced by a better shopping experience and more transactional efficiency.
Item recognition for non-grocery items
Smart vision technology is also applicable in areas other than fresh produce. A great example can be found at furniture retailer IKEA, which uses camera-equipped self-checkout solutions. A camera mounted on top of the self-checkout device detects any scan errors and notifies the customer to rescan and correct the shopping basket if necessary. IKEA has deployed this automatic item recognition technology in several countries in Europe, a camera-based self-checkout solutions empowered by an open-API software platform. The technology improves the self-checkout experience allowing consumers to scan and go, thus avoiding queues or other frictions. It is an example of how smart vision technology based on AI can improve the customer experience by eliminating frictions and reducing queues. Not a surprise,
IKEA was awarded the "Retail Technology Award Europe 2021" in the category of Best AI & Robotics Application.
Want to learn more about other smart innovations in self-checkout,
download our whitepaper.
HANDEL
