Grocery Store Implements Fresh Produce Recognition Technology

While AI-driven in-store technologies are on the rise due to the pandemic, developing and implementing fresh product recognition and maintenance systems is the central problem that companies are trying to solve. An EU-based startup Pixevia is beginning to offer European SMEs a fully tested version of this technology.

The COVID-19 pandemic and its disrupting effect on the retail sector have accelerated some in-store technology trends like biometric checkout and scan-and-go systems. The new trend that is emerging now is the drive to look for solutions to distinguish fresh produce—like fruits and vegetables—as they tend to perish fast, and basic AI systems are not always designed for such fine distinctions.

In the U.S., companies like Amazon are responsible for pioneering many AI-based automation solutions. For example, the company behind the first cashierless shops has recently launched the Amazon Fresh store chain with a bigger grocery offering and a greater emphasis on product freshness. These stores utilize AI-based grocery product recognition technology that eliminates the need to scan each perishable item individually, reduces the check-out hassle, and minimizes customer contacts.

Relevance of such in-store solutions is also increasing in Europe. PIXEVIA—a store automation technology provider whose product recognition system now powers Europe’s first cashierless grocery store—is one of the pioneers in the field.

“There are many challenges involved in developing a fully functional and robust grocery product recognition system, especially when we’re dealing with ‘fresh’ category goods,” said Mindaugas Eglinskas, CEO of PIXEVIA.  “For example, many algorithms fail to distinguish between similar-looking fruits—such as oranges and persimmons. Also, since most fruits and vegetables have a very limited shelf-life, there is always the risk of failing to control the food inventory and creating a lot of food waste. Therefore, a market-ready system has to track shopping trends, generate precise sales forecasts, and issue discounts autonomously. Given these challenges, we have tested our product meticulously in our PIXEVIA Smart Store in Vilnius Old Town and have arrived at a fully functional grocery recognition system.”

While the primary focus for PIXEVIA team was initiating PIXEVIA Smart Stores, designed to make shopping for essential goods easy, quick, and convenient, over time, the interest to develop transferable AI technology for retail locations primarily bigger than 150 sqm across Europe became the main focus. With its wide range of functionality, PIXEVIA’s grocery product recognition technology is one of the company’s key products.

“Our AI-based system successfully recognizes different fruits, salads, bakery goods as well as meat and dairy products, allowing customers to forgo the long and often arduous check-out process. In addition to that, it is continuously monitoring product freshness. In case a product is approaching its expiration date or showing any signs of staleness, the dynamic pricing algorithm alerts the store manager that the product has to be sold at a lower price and later if necessary taken out of stock,” commented Mr. Eglinskas.

According to Mr. Eglinskas, the system issues discounts by analyzing the data pertaining to the previous consumption trends as well as weather patterns. This way, the in-stock goods are maintained fresh and safe to consume, as well as sustainable consumption is secured.

With the rapidly increasing Amazon Fresh locations, it appears that the U.S. is at the forefront of AI-based in-store system adoption. However, smaller companies are making significant progress in accelerating the spread of this technology in other regions. In Europe, PIXEVIA takes a leadership role in this field.