mobile background

AI Retail store


AI Retail store

A single Deep Runner identifies customer movements through multiple CCTV cameras installed in the retail store. Tracking is possible even when a customer moves from one camera to another. This device records the number of customers entering and exiting for each specified time, and it is possible to determine which products the customer is most interested in, and which store space has high or low efficiency.

Such a function provides big data on the movement path of people, and can be used, for example, to improve usability by identifying and analyzing the movement and frequency of use of public facilities users.

Applications

  • Gather marketing data for retail store
  • Identification and improvement of movement lines around public facilities
< Visitor movement analysis >
< Visitor movement analysis >
< In-store hot-zone analysis >
< In-store hot-zone analysis >

Configuration

Input/output

  • Remote control functions

            Camera RTSP Address Mapping

            Setting counting entrance

            Registration of excluded areas


  • Real-time output

            Coordinate information of each customer's movement

            Unique hash code for each customer

Features

  • It uses the latest deep learning technologies, pedestrian detection and person re-idenfication, to recognize and track the customers.


  • Store employees in certain uniforms may be excluded from analysis.


  • With the built-in counting function, customers entering and exiting a specific entrance can be counted.


  • A single 12 x 14 cm compact Deep Runner Z5R board can be simultaneously recognized by up to 16 high-definition CCTV cameras. Built-in video decoder eliminates the need for separate devices.


  • As real-time recognition is performed from IP cameras, existing CCTV cameras installed in the stores can be used.


  • The built-in web-based interworking protocol allows easy interworking with IT servers.