Colruyt Lowest Prices is the first Belgian supermarket to test automatic product recognition. In the store at Kortrijk, fruit and vegetables are automatically recognised at the checkout by means of artificial intelligence. In collaboration with Robovision, a smart camera was developed and installed above the scales immediately determining the product concerned. This test takes Colruyt another step ahead in digital progress. For three months, co-workers and customers of the Colruyt store in Kortrijk will work together to find out how this technology can reduce time and increase efficiency at the checkout.
How does it work?
Colruyt is the first Belgian retailer to test the automatic recognition of fruit and vegetables. "Together with our customers from Kortrijk, we will examine for three months whether this new technology makes the checkout process more efficient", says Colruyt regional sales manager Rudi Dewulf.
When customers put a vegetable or fruit in the scales at the checkout, the camera takes a picture. Based on this picture, the system can automatically recognise the product in the scale. Customers then use a tablet to confirm the correct product.
The software recognises 120 kinds of fruit and vegetables.
In collaboration with AI business Robovision from Ghent, a unique 'deep learning' model was developed that can recognise 120 kinds of fruit and vegetables at Colruyt. Thanks to artificial intelligence, product recognition becomes more accurate as more pictures are added. The first tests revealed that in 97% of the cases, the correct product is presented on the tablet. "We have customised the software for Colruyt”, says Jonathan Berte, CEO of Robovision. "It is nice to see that our expertise in AI found its way to the store. This is a unique partnership with an innovative retailer. Moreover, the project is a perfect match with our mission: making AI more democratic."
Colruyt is testing automatic product recognition to find out whether it can make the checkout process more efficient. We consciously chose fruit and vegetables because it is a product range without bar codes. Until now, our sales employees had to type a code for all fruit and vegetables at the checkout; this is no longer necessary with automatic product recognition. Secondly, the sales employee can scan other products while the customer weighs his unpacked fruit and vegetables with the camera. The result may be that we have to pack much less fruit and vegetables in the future. "Today, we pack some fruit and vegetables to avoid confusion between products. If we can automatically recognise them, this problem might vanish", says Rudi Dewulf.
We deliberately chose Colruyt Kortrijk as a test store, because it is one of the pilots where we only offer reusable bags for fruit and vegetables. The software is able to subtract the weight of the bag from the weight of the fruit and vegetables.
The first test showed that customers react enthusiastically and are pleased to use this new technology because they can take an active role in the checkout process. Colruyt chooses to maintain the classic way of weighing by the store employee during the test in Kortrijk. Customers can choose between the two systems.
Colruyt is constantly investing in new digital developments to make shopping more efficient. Colruyt thus introduced its electronic price tags this year, organised a speech technology test in February and is continuously expanding its MyColruyt app. "We believe that product recognition can replace the bar code, which was also innovative 30 years ago."
Extending product recognition to other products is not yet planned. "We are curious and waiting for the test results to find out what this innovative technology can mean for the future", concludes Rudi Dewulf.
Download your pictures here.
Press release distributed by Pressat on behalf of Pressat Wire, on Thursday 8 August, 2019. For more information subscribe and follow https://pressat.co.uk/