Watch CNN's Coverage
of VideoMining

VideoMining Behavior Engines are real-time software modules that automatically recognize over 100 types of human behavior inside a store. The engines work on PC hardware and utilize standard security cameras. The VideoMining Behavior Engines are based on over a decade of groundbreaking research and patent-protected technologies for video analysis and provide an unprecedented visibility into what is happening in your store, giving retailers The power to see more

The five VideoMining Behavior Engines provide different sets of real-time measurements that are most valuable to retailers.

Traffic

The Traffic Engine provides best-in-class counting and additional unique features by using sophisticated person detection and object recognition technologies (allowing filtering of shopping carts and identification of shopping groups). The traffic engine works with standard security cameras to extract the following information:
 

  • Accurate store entrance and exit counts
  • Traffic counts for individual store categories
  • Real time store population count
  • Real time category population count
  • Traffic dispersion (initial direction upon entry)
  • Estimate of shopping groups

Queue

The Queue Engine measures key properties of the queue and derives information regarding the interaction of customers with customer service points in a store, such as the checkout and information desk. Key measures include:
 

  • Queue length
  • Wait times
  • Service times
  • Queue occupancy (% full)
  • Queue overflow detection
  • Queue abandonment rate
  • Unattended service point
  • Number of employees at service point

Interactions

The Interactions Engine detects and measures various interaction behaviors, such as shopping a category, browsing a particular shelf and interacting with a specific product. Sample data includes:
 

  • Number of visitors shopping a given category
  • Number of visitors browsing a particular aisle
  • Number of visitors interacting with a particular fixture/product
  • Length of time spent shopping, browsing or interacting
  • Visualization of relative shopping activity in an area
  • Changes in shopping activity over time
  • Path analysis in area (dominant directions and speed)

Impressions

The Impressions Engine measures customer response to specific in-store marketing elements, such as static signage, POP displays and other forms of digital media, including retail TV networks and kiosks. The engine uses unobtrusive cameras and sophisticated face analysis to extract the following types of data:
 

  • Impression count (visitors who actually viewed media)
  • Average length of impression
  • Distribution of impressions by time of day
  • Rating of media effectiveness based on audience response

Trip

The Trip Engine uses VideoMining’s breakthrough technology for tracking a person across multiple cameras to analyze the entire shopping trip of store visitors. The engine extracts key measurements along the entire shopping trip (from entrance to checkout) that deliver deeper insights about the trip as a whole. The derived insights include:
 

  • Traffic distribution map (shows penetration throughout store)
  • Shopping distribution map (shows activity throughout store)
  • Shopping conversion map (shows traffic converted to shopping)
  • Category-level traffic distribution map
  • Category-level shopping distribution map
  • Category-level shopping conversion map
  • Shopping sequence analysis (order of categories shopped)
  • Category-level path analysis