Smart Video Analytics Platform Delivers Real-time Crowd Intelligence

WaitTime intelligent software deciphers crowd behavior in a whole new way. The proprietary AI technology assigns behavior profiles based on movement and speed, unlike other existing crowd technology that relies on Bluetooth connections or beacons. Here’s how WaitTime keeps crowd data anonymous–and is changing the waiting game.

Image credit: WaitTime

Deciphering recorded video footage to optimize business operations is nothing new, but industries that host large or diverse crowds often need extra help. Companies catering to the large groups of people that circulate stadiums, malls, and other public venues mostly still rely on human observation of video footage to “intelligently” provide behavior analysis. Others use beacons or Bluetooth to track behavior.

Securing accurate and timely crowd behavior analysis is vital to everyone’s safety. If one area gets too congested often, there could be fire hazards or other consumer health risks from massive gatherings that aren’t managed properly. Soccer matches and concerts are events that come to mind where crowds have exhibited dangerous behaviors.

Gorilla Technology Group, a company that focuses on video, intelligence, and IoT technology, made waves in retail by providing a smarter way to gather and analyze crowd behavior data. An Intel® IoT Solutions Partner, Gorilla’s Smart Retail Solution involves IoT sensors and video cameras that monitor crowd behavior via connection to Bluetooth.

Gorilla Technology’s Smart Retail solution uses its Intelligent Video Analytics Recorder (IVAR) edge AI platform to collect and correlate operational, shopper, and product data for retail business owners. Smart Retail uses cameras, IoT devices, and machine learning to determine the number of shoppers in a store and help shape sales, marketing, and advertising strategies.

Smart Retail can monitor staff and assess the store layout. It detects how shoppers navigate through the store and analyzes dwell time to determine consumer behavior and conversion rates, along with product and brand popularity. Store owners can use that data to deliver targeted advertising, adjust staffing needs, and alter inventory to meet market demands.

Using Crowd Movement to Optimize Ops

Building on the growing need for crowd intelligence in retail and beyond, one data-driven company is pushing through crowd behavior analytics with intelligent software that observes and differentiates guest profiles through varying types of movement. Based in Detroit, video analytics company WaitTime has brought crowd behavior analysis to a new level. More specifically, it has patented multiple algorithms that analyze crowd behavior based on the speed and direction of movement of each person, 24 times per second, explains Zachary Klima, Founder, and CEO of WaitTime.

Once WaitTime’s IoT sensors are connected and placed on video cameras, the internal software uses footage data and AI to determine which guests are pedestrians passing through the area, congregating off to the side, or guests specifically in line and advancing to the front. The AI software groups varying movements into defined behavior profiles that operators can use to determine what visitors do inside and around their establishments, all in real time.

Fueled by Intel® Xeon® Scalable processors with AI accelerators in combination with Cisco’s Meraki and analytics solutions, WaitTime uses AI to precisely count the number of people in a given area and determine whether they are congregating or passing through. WaitTime also provides detailed analytics that operators can use to provide better shopper experiences.

Different use images of WaitTime technology.

Image credit: WaitTime

Smart Pixel Tracking with Algorithms

WaitTime developed four unique algorithms that analyze crowd movement with developed pixel tracking technology. Each algorithm is crafted for solutions that fit varying sports and entertainment venues, airports, retail malls, exhibition centers, and casinos.

1. Queueing Algorithm is an API output of the number of people in line and the rate-of-service. This observes movements from an overhead view and calculates how long it takes for a person at the back of the line to reach the front. This algorithm specializes in detecting unstructured and free flowing lines at concession stands, product areas, or ingress gates.

2. Stanchion Algorithm is similar to the Queuing Algorithm, but it’s specifically for structured lines at cafes, betting lines, or security sections such as TSA at an airport.

3. Massing Algorithm provides a side view of crowds and calculates occupancy density of a given area, from zero to 100 percent. The software still detects movement through pixel tracking but uses a heat-map style. It can intelligently determine density based on moving pixels that are grouped together as a crowd would be. This is good for areas within convention centers, retail spots such as checkout counters, and smart campuses.

4. Entry Exit Software monitors how many people enter or exit an environment in real time. This is helpful for clubs, offices, casinos, and malls that may risk reaching occupancy limits.

Operators and guests can both benefit from WaitTime algorithms. Businesses and venue managers can access all their data via an interactive WaitTime crowd management reporting tool, where they can observe a complete picture of everything happening within their establishment as it happens. They can access live footage, reports, and analytics—all easily exportable into CSV or other formats. WaitTime also partners with InstantConnect to create dedicated voice channels or dynamic talk groups, to dispatch frontline teams to quickly solve issues the moment they occur.

On the consumer side, guests can access the application to uncover actual wait times at specific concession stands. The valuable info is generated from crowd behavior data and video footage. For instance, WaitTime recently partnered with the Denver Broncos. Guests attending Broncos games at Empower Stadium can now use the WaitTime application to see which concessions have the shortest or fastest lines.

How does WaitTime differentiate itself from competitors? “We are very different because we don’t track Bluetooth, cellphones, Wi-Fi, or any of that,” explains Klima. “We are tracking actual movement of the human body itself–anonymously. This is a very detailed pixel tracking technology, which is how we can get such high accuracy with our software.”