Fix Price, supported by Neoflex, connected 6,000 stores of the chain to the video analytics system

2024: Implementation of the video analytics system

Fix Price and Neoflex have implemented a full-scale video analytics system using computer vision based on the Yandex Cloud platform. The solution has already accelerated customer service and allowed traffic to increase in more than 6,000 stores in the chain. Fix Price announced this on July 3, 2024.

Earlier, Fix Price did not have a system that analyzes the video stream coming from already equipped cameras video surveillances in the chain's stores. The Neoflex team implemented a server Cloud computing solution Yandex that made it possible to quickly identify queues at cash desks and inform store staff about this.

To implement the project, a system was developed that analyzes images obtained from CCTV cameras installed in stores using a computer vision model. The solution allows, in a mode as close as possible to the real time, to determine stores in which the queue for the cash register exceeds the established limit on the number of customers. In addition, the store manager receives a notification about the need to open an additional cash register. The development and operation of the system is implemented in Yandex Cloud on a serverless architecture, which made it possible to significantly reduce the time for the implementation of the project and reduce the costs associated with the infrastructure during the piloting and replication stage through the Pay-as-you-go approach.

The key task that our team faced was to develop a universal solution without the need to re-equip the entire chain of stores. Neoflex's expertise in Data Science and computer vision, as well as the use of Yandex Cloud services, made it possible to get the first version of the model and start piloting after 1.5 months. Thanks to the high involvement of the project team on the part of Fix Price, which promptly adjusted the angle of view of the cameras and provided the necessary information, it was possible to successfully achieve the set goals on time, "said Denis Knyazev, director of business development at Neoflex.

Measurements of the video analytics pilot project in December showed an increase in point turnover by 2% as a result of the introduction of the technology compared to comparable stores where such technology was not used. In addition to queue control, we plan to expand the functionality of the system for monitoring the display of goods and the state of the trading floor. The system will track "empty shelves," monitor the cleanliness of the floor and the absence of foreign objects in the aisles, which will help improve the purchasing experience, - commented Oleg Lexin, head of IT Fix Price.