New Neural Network Trained to Assess Customer Activity in Stores

A master’s student of the St. Petersburg Electrotechnical University has developed an intelligent system that analyzes the movement of customers in stores. The programme can be used to determine popular products among various social groups, as well as assess the effectiveness of personnel, advertising, and promotions for goods, TASS reports. 

It is expected that the project will allow analyzing the shopping activity in the store and, based on the data obtained, improve the marketing strategy of a particular company. The system will identify the most popular products for various groups of visitors and compile statistics to assess the effectiveness of advertising.

In addition, the development will make it possible to determine the optimal number of consultants and cash registers to provide quality service and balance the workload on staff. During an evacuation, the system will be able to provide data on the number of visitors, as well as determine their age and gender.

To identify people on video and recognize their faces, specially trained neural networks will be used with the methods of projective geometry (the image obtained from CCTV cameras is converted into a top view). Also, currently, the work is underway to apply such a system to control social distance in the fight against the spread of coronavirus.

Ru-Main, 18.11.2020 

Related Posts

Leave a Reply

Your email address will not be published. Required fields are marked *