09 Nov Algorithms in action
Timothy Compston discovers why retailers have been counting on video analytics to keep their businesses on track
When it comes to the state-of-play of video analytics (video content analysis) in the retail space there is little doubt that analytics has proven invaluable during the ongoing coronavirus 2019 (COVID-19) pandemic. Capabilities like people counting; face mask detection; and proximity detection algorithms – to support social distancing – have all been front and centre of efforts to return commercial activity to a semblance of normality.
Certainly, retail is no stranger to video analytics from a security, loss prevention, and business generation perspective. The new and refined capabilities being developed for COVID-19 are a welcome addition to well proven solutions like heat maps, to see which areas in-store are gaining the most traction, and queue monitoring and management. Significantly, they also join other smart innovations that have been developed in recent years to address pressing retail issues. A case in point being StopLift Checkout Vision Systems (now part of NCR Corporation), and the advanced AI (Artificial Intelligence) driven computer vision algorithms employed to work with in-store cameras – and cameras built into the checkout systems themselves – to help distinguish between legitimate and fraudulent behaviour at manned and, increasingly, self-service checkout areas.Moving ahead it will be interesting to see which of the pandemic-specific features of video analytics that vendors have sought to unlock to help retailers comply with new temporary regulatory requirements are retained in-store once COVID-19 is brought firmly under control.
Fast-track solutions
Given the current pressures on retailers, it is perhaps not too surprising that numerous vendors have been rushing to bring new video analytics-related solutions to market. A case-in-point is Senstar which this month announced a Safe Spaces video analytics solution in a compact appliance, the Senstar Edge Platform. The chief operating officer at Senstar, Jeremy Weese, stresses that the company has opted to deliver a complete solution, operating on an edge platform, so it is affordable and easy-to-use, while being VMS and camera agnostic: “The solution includes face mask detection, physical distance monitoring, occupancy counting, and sanitisation station monitoring analytics,” concludes Weese.
Rewinding to June and March Networks unveiled its own answer to the ever stricter COVID-19 requirements with a Health Compliance solution encompassing a suite of new features including, crucially, real-time occupancy management. The solution, according to March Networks, is based on the vendor’s data analytic platform – Searchlight – and, consequently, seeks to combine data from video surveillance cameras and third-party sensors to alert operators of potential compliance violations.
Drilling down into the real-time occupancy element of March Network’s Health Compliance offering, using Searchlight, retailers can count the number of people entering their locations in real-time and visualise occupancy data in multiple ways, both locally and from a central location. March Networks says that occupancy alerts may be displayed on a local, colour-coded video client or public view monitor, a mobile phone or tablet or via a smart lighting system, which can change colour to indicate when a retail store is nearing or has reached its maximum capacity. Already, March Networks reports that the new solution has been selected for roll-out by a major European supermarket chain with more than 150 locations.
Real-time occupancy
Seeking out the thoughts of Trevor Sinden, director of sales at March Networks for the Middle East & Africa, on what retailers are looking for in this new COVID-19 reality, including with regards to people counting, he explains that the vendor has seen a growing requirement for solutions for real-time occupancy management and remote capabilities that allow retailers to monitor their businesses from afar in addition to their regular security requirements: “Real-time alerts, in particular, when combined with video surveillance, are of great interest to our customers. This can ensure that businesses are instantly notified when there are issues with health compliance, such as when they are close to their occupancy threshold or when it has been exceeded.”
Expanding on the rationale for the release of the Health Compliance solution, Sinden adds that when the pandemic hit March Networks knew businesses would need a solution that would allow them to continue to operate while ensuring they could do so safely: “March Networks has introduced the Health Compliance Solution to help retailers protect their employees and customers with real-time alerts during COVID-19. Built upon our Searchlight video-based business intelligence platform, this solution can measure building occupancy, rapidly detect individuals who may have elevated body temperatures, and it also ensures compliance with new cleaning and sanitisation procedures through convenient video audits. It also has the capability to integrate with smart hue lightbulbs, which can be installed at entrances and exits for another visual indicator of occupancy.”
Talking about what is driving the retail market in the Middle East region, Sinden notes that retailers here have started to recognise the importance of business analytics in their stores combined with video security: “We see many retailers looking for AI [Artificial Intelligence] based solutions to increase their profitability by reducing shrinkage. Many are looking for ways to address POS [Point of Sale] fraud and improve inventory management. We have also seen instances where retailers have been fined, and in some cases had to close, due to breaking social distancing regulations so solutions that can help with COVID-19 compliance are important.”
App challenge
Of course vendors of all sizes are working on solutions to address the latest challenges for retailers as evidenced by initiatives like Security and Safety Things (S&ST) GmbH – a subsidiary of Robert Bosch GmbH – that operates an open IoT platform that connects camera manufacturers, SoC (System-on-a-Chip) providers, application developers, VMS providers, and systems integrators. Back in August, Security & Safety Things announced that 21 innovative applications are in development as part of its App Challenge to create new solutions for the retail, transportation, and manufacturing sectors.
Focusing on the retail sector specifically, S&ST reports that apps are in the works – that can run on cameras powered by the operating system developed by Security & Safety Things – to track occupancy levels and detect facial coverings, perform heat mapping or customer traffic flow, assist in preventing shoplifting or address COVID-19 specific challenges such as the detection of proper face protection, occupancy management, and adherence to social distancing. Alexander Harlass, head of developer relations at Security & Safety Things (S&ST) is pleased with the number of participants and the variety of applications designed to assist retailers that have been entered for the first App Challenge: “Like many industries, retailers are facing a unique set of challenges in regard to providing a safe, hygienic environment for both shoppers and employees.”
Speaking to a selection of the app developers about their solutions and the current retail landscape, Alessia Saggese, sales and marketing director at A.I. Tech – which is responsible for the AI-CROWD-DEEP and +AI-FACEDETECT-DEEP apps – notes that today in-store experiences are heavily influenced by a store’s ability to ensure that social distancing can be adhered to: “To assist, A.I. Tech saw the opportunity to develop AI-CROWD-DEEP for occupancy management and +AI-FACEDETECT-DEEP, which detects the presence of facial coverings, to ensure the safety of staff and customers at all times, while also maximizing the in-store customer experience to encourage a return to the physical shopping experience.” Saggese recognises that the absence of proper facial protection is a real area of concern for retailers: “With that in mind, A.I. Tech developed the +AI-FACEDETECT-DEEP application for detecting and generating an alarm when someone enters without a face mask.”
For his part, Moritz Prinz, partner, TNG Technical Consulting GmbH believes that the COVID-19 crisis has necessitated new and unconventional solutions and, crucially, was why the developer participated in the Security & Safety Things (S&ST) App Challenge: “Particularly in retail environments, conventional feedback terminals are continuously touched by many people, and potentially spread the virus, therefore we came up with a contact-less solution.” He stresses that with the hand gesture detection Visual Feedback app customers can now give contactless feedback without having to touch any kiosk or device: “It can also increase the amount of feedback retailers receive, while showing they are adhering to hygiene policies and dedicated to the well-being of customers.”
Significantly, Prinz adds that the flexible infrastructure of the S&ST platform meant that it only took four months to put the idea for the hand gesture detection Visual Feedback app into practice: “This application received the highest rating in the S&ST App Challenge for the Retail and Commercial Buildings category.” Alongside this, Moritz says that a Dynamic Privacy Mask application has been developed by TNG Technical Consulting that can also be deployed in a retail environment.
Turning to Tijmen Vos, technical director at Eagle Eye Networks – which offers a social distancing app – he explains that, using existing video streams, Eagle Eye Networks is capable of delivering high quality forecasting, analytics and management tools to help retail entrepreneurs improve the business side of their operations as well as focusing on a safe shopping experience: “Distance management has become a social responsibility and using innovative technology is a great way to make sure that retail shops can stay open while keeping the required safety distance.”
Changing behaviour
On the question of how the COVID-19 crisis has influenced the way video analytics is being applied by retailers and the behaviour of shoppers today, Vos, from Eagle Eye Networks, reckons that COVID-19 has caused a few permanent changes in the retail sector: “For example, research has shown that supermarket visitors have changed their shopping patterns to alternate between days, times, and frequency, mainly to avoid large crowds during rush hours to add an additional layer of safety. Customers have become more cautious and aware when it comes to safety and security while being in a public environment.” Vos adds that Eagle Eye Networks believes that video solutions with analytic tools, such as customer count or line crossing, could be a way to drive change and deliver retailers helpful tools to improve customer management as well as making the shopping experience as safe as possible for customers and employees.
Pull quote: “Like many industries, retailers are facing a unique set of challenges in regard to providing a safe, hygienic environment for both shoppers and employees.”