Dutch firm Sightcorp has developed a smart store traffic light system that combines AI and digital signage to make entrance flow management in shops manageable and safe in Covid-19 times.
 
During the course of the pandemic, retailers and other public venues have been hacking together communication around safety measures and guidelines leaving shoppers feeling unsafe and uninformed. The impact of Covid-19 has changed how people go about their daily business, heightening awareness of germs and contagions and making people reconsider all the surfaces they encounter.
 
As the lockdowns begin to ease and a return to normality is within grasp, core public infrastructures such as businesses, transportation and retail are recognising the need to adopt a consistent approach to providing services with Covid-19 regulations in mind. Technology can be used as a flexible and effective tool to address this problem and to prevent the rise of new infections.
 
Sightcorp’s technology builds on its DeepSight crowd analysis software. It combines sensors, AI software and digital signage to create a smart store traffic light system enabling businesses to have a safe environment as people make their way back to stores.
 
The technology adopts a multi-faceted approach using facial analysis and crowd control to create an entrance flow management system. The software performs an anonymised analysis on an individual or crowd basis. No personal data are collected or stored, which makes this privacy proof. The gathered data can be integrated with a content management system or a reporting dashboard for analysis of trends.
 
The technology can be applied using digital displays, which can switch between displaying targeted advertisements to safety communications and access control updates. The software can determine whether an establishment has reached its full capacity, using a people counting system at the entrance and exit.
 
From there, key messages can be transmitted to customers arriving at an entrance via a digital screen. The screen content can be programmed to display real-time updates on whether it is possible to enter the store, while also including a stop-and-go traffic light system to ensure complete message transparency. This contactless method of controlling the flow of traffic into premises could provide relieve staff members of queue management tasks.
 
The system comes with a face-mask detection algorithm that can be turned on and off when needed, allowing for real-time mask detection of shoppers entering the store. If a visitor is found not to be wearing a mask or wearing one incorrectly, a reminder message can be displayed on the screen, or a trigger can be sent to an access control gate or a smart device such as a smartphone or smart watch.
 
Computer vision technologies are becoming widely used and incorporated in these times to aid in safe reintegration back into society. They have the potential to become part of public infrastructure and, when used effectively, can make communities safer and lives more convenient, and become a resource for security. The impact of Covid-19 has been significant but having these products available can help make it easier to cope with future pandemics.
 
DeepSight lets adjustments be made as circumstances evolve or as any new health threats emerge for the public in the future.
 
“I truly believe that the newest addition to our already powerful audience analytics product can aid retail sectors across the globe to communicate the daily changes in rules and guidelines in a smarter and effective way while keeping everyone safe,” said Joyce Caradonna, CEO of Sightcorp.
 
Apart from detecting masks, many retailers and advertisers use audience measurement software for people counting, demographic analysis and for gathering advert performance metrics. Therefore it’s important for the technology to remain robust and accurate even when analysing faces partially covered by face masks. DeepSight models have been developed and re-trained with this in mind and are still able to detect and estimate the gender and the age of all audiences. The audience measurement and face masks detection technologies run in parallel so retailers can make sure their visitors wear face masks during shopping while still collecting demographic data.
 
Sightcorp is an AI spin-off from the University of Amsterdam specialising in face analysis and face recognition software. Through computer vision and deep learning, it can analyse and recognise faces in images and videos, and in real time.