X5 to roll out computer vision technology

Moscow, 6 June 2018 – X5 Retail Group ("X5" or the "Company"), a leading Russian food retailer (LSE and MOEX ticker: FIVE), has tested video analytics and computer vision technology based on neural networks and AI. The technology contributes to a tenfold increase in the speed of the planogram control process, while also reducing the number of people leaving stores without a purchase by 10% and shrinkage levels by 20%. The technology can be leveraged to boost sales, especially when deployed in real time.

Testing revealed that video analytics add most value to X5 stores when used to control layout and availability of products on shelves, monitor queue length and prevent store thefts, so the Company will focus on these application areas.

Intelligence Retail, a solution developed by the company of the same name that is a resident at the Skolkovo Foundation's IT cluster, scored a 93.7% accuracy rate for on-shelf product recognition. It was tested in five Perekrestok stores in the Moscow Region, notifying employees of empty shelves and the need to replenish the products. During the pilot stage, the technology expanded its recognition capabilities to some 1,500 products. A similar technology was tested at the Pyaterochka chain. Photo and video control over the layout proved efficient, boosting sales of target products.

Video analytics also help to improve the quality of checkout services. Technology for counting the number of customers in a queue is now being rolled out at Karusel hypermarkets and tested in 20 Perekrestok stores in Moscow. When there are more than five people in a queue, store management gets notified of the need to open an additional checkout lane.

This year, X5 plans to scale pilot testing of selected technologies to dozens of stores and integrate computer vision into its business processes and IT systems. The Company estimates that about 60% of its stores can greatly benefit from available video analytics and computer vision tools.

Download (PDF. 0.2 MB)

contact

Send us your feedback