AI based service that provides store information by counting the number of tables and customers in a store.
AI based service that provides store information by counting the number of tables and customers in a store.

FUNCTION

  • Use Body Detection to determine number of customers and free tables in the store
  • Provides information on the ‘crowdedness’ of a store, and table information
  • API Offerings
    • Total number of customers
    • Total number of tables
    • Number of free tables
  • Use Body Detection to determine number of customers and free tables in the store
  • Provides information on the ‘crowdedness’ of a store, and table information
  • API Offerings
    • Total number of customers
    • Total number of tables
    • Number of free tables

Configuring Unoccupied Table Space

  • The Manager will be able to designate table areas from a video feed.
  • The photo shows 5 designated table areas, the green pin indicates that the table region is unoccupied.

Count Table and Number of Customers

  • Counts when a person sits in the table area.
  • The two yellow pins above the tables indicates that two tables are in use.
  • Three of the five tables are empty.
  • 3 Customers are detected, with 2/5 tables occupied, hence indicating that the store is ‘uncrowded’

TECHNOLOGIES

Object Detection (Human Detection, Table Detection)

Separation and analysis of object regions

Inter-object correlation analysis

SERVICE

1. The manager of the store designates the location of the table in the photo.

2. The IP camera in the store transmits photos at a fixed cycle (FTP transmission)

  • Video transmission period can be set from 5 seconds to 5 minutes

3. The photograph transmitted from the in-store camera is analyzed, and the total number of customers, the number of tables in use, and the number of free tables are available through an API.

FEATURES

  • Deep Learning based Human Detection enables accurate counting.
  • Easily specify table location with an intuitive drag-to-select UI to determine total table count and location
  • Can be used to monitor crowdedness
  • Only still photos are transmitted over the network, not the full video stream, saving costs