08 May Adding analytics to AI
High value applications for AI video analytics are emerging at an increasing rate, and are not limited to security alone, writes Dennis Choi, General Manager, IDIS Middle East & Africa.
Video analytics, powered by AI, have been transforming security operations with ever more reliable tools for detection and verification, automating many of the system monitoring and review functions that previously required manual input from security staff. The trend towards increased reliability, power, and affordability is continuing, with a growing range of add-on analytics tools making AI video easier and more affordable to adopt. Alongside this, we are seeing the emergence of applications beyond security. These new uses are adding further value to surveillance systems, and underscoring business cases for upgrades and new installations.
One sign that the video analytics market is really maturing is that end-users themselves are finding new applications and discovering new uses. From logistics and retail to banking, hospitality and critical infrastructure, the more AI video becomes embedded in operations, the more useful it is proving to be. The benefits are potentially far reaching.
Addressing the pain points
Wider and more ambitious deployment of AI video, targeting surveillance for novel purposes, involves security leaders in alliance-building across their organisations. Through these projects they are engaging with the priorities and pain-points of different departments that can potentially benefits from AI – for example, the customer service directors who want to reduce queuing and waiting times; the operations directors who want to speed up deliveries or optimise staff deployment; or the marketing directors who want a better understanding of customer movements around merchandising and special promotions. This trend aligns with a more general drive to digitisation, a pattern that we are seeing across economies as organisations look to achieve efficiency gains by automating and adding AI-capabilities to systems from payroll to helplines, from delivery processing to marketing. As this continues, it is bringing the benefits of greater stakeholder engagement and advocacy, raising the profile of security departments, and unlocking bigger budgets. We are seeing increased focus on health and safety across many sectors. Intelligent video analytics tools such as line-cross detection linked to automated PA reminders, can provide an extra layer of protection in a wide range of settings. They are being used to enhance surveillance in locations where people are known to stray and where they are at greater risk, despite warning signs and other measures – leaning over safety barriers for example, standing close to rail platform edges, or crossing into restricted zones where heavy machinery operates.
The same automated detection capability can underpin more effective and rapid intervention by staff, including security officers, if people disregard warnings –and there are several reasons why they might, from complacency, and lack of due attention, a physical disability or mental vulnerability. We see many applications for AI-video in sterile environments: in food processing, drug manufacturing, healthcare settings, laboratories, and research facilities. In these settings, tools such as mask-wearing detection that proved their value widely during the pandemic, are now finding routine use helping to enforce hygiene protocols. The region’s banking landscape has been shifting up a gear, with both retail and corporate banks recognising the need to invest in digitisation and analytics, and to embrace the opportunities of tech in order to remain competitive. Banks are also responding to wider changes, including the diversification of the region’s economies away from oil, and are looking for new opportunities, for example with extended personal loan portfolios putting more emphasis on mortgages and lending to SMEs; and with efficiency savings through a continuing reduction in the number of branches and exponential increased use of ATMs. Retail banks have been leading the way, and corporate banking is not far behind. Both are breaking new ground in their use of video analytics functions including queue management tools, heatmapping, and occupancy monitoring.
Multitude of applications
Similarly, they are using AI for car park management, with authorised lists for staff parking, and line-cross safety demarcation around non-parking zones and driver blind-spots, both improving safety. Analytics are enhancing protection against theft of and from vehicles; and they are enabling greater vigilance when pedestrians enter parking areas, for example allowing anti-social behaviour to be spotted more reliably. AI video has flexibility built-in, so the different risk profiles of branches in different locations can be accounted for. So, a line-cross event that triggers a priority alert at a remote or higher risk location, will only set off a routine notification at a lower risk site, or a city centre location where pedestrians are expected to be as a matter of routine. In logistics settings, and manufacturing, companies face continuing pressures on supply chains, and there is much greater awareness of the need to design-in operational resilience. Digitisation, including video analytics, is one of the key enablers. And we are seeing new value being gained from existing surveillance infrastructure through affordable analytics upgrades and add-ons, as well as investment incompletely new systems. Cameras that are already in place, for security and safety surveillance, are being enhanced with AI so that they can now do new jobs and deliver additional, high value benefits. Examples include heatmapping being used to identify and monitor bottlenecks and cold spots, and to measure areas of activity against minute-by-minute process and production schedules. This is enabling both more efficient management of operations in-the-moment, and better planning longer term as processes and resourcing are refined. Analytics are increasingly being trusted with compliance monitoring, for example preventing fire risks from going unnoticed– fire extinguishers being removed, objects blocking exits or escape routes, or doors being propped open. And object detection can spot when essential equipment is not in its correct place – an endless list of potential breaches that can reduce efficiency, or present safety risks. If the presence of something is critical, analytics can be used to monitor it – a fact that is increasingly understood, and is leading to new applications. In warehouses and industrial settings where safety adherence is critical, some of the most useful recent advances in video analytics include tools for fall detection. These are helping to mitigate one of the biggest risks in these settings globally: slips, trips, and falls
Finally, the retail sector across the Middle East is a significant growth area of AI deep learning tools. Retailers are keen to leverage greater value from their existing, surveillance investments, and want to take advantage of the business intelligence that systems can deliver when upgraded with analytics functions –insights around customer behaviour, staff interactions, point of sale efficiency etc. In Turkey for example, where we opened our new Ankara office last year, we are working with two major fashion brands to upgrade over 70 stores, adding in exactly this kind of retail analytics functionality. This is a significant potential growth market, with European and U.S. brands continuing to expand into the Middle East and diversify manufacturing and distribution operations away from China and Bangladesh to countries such as Turkey and Jordan, as they look to reduce supply chain risks. The security sector is well placed to help these markets develop, with tools that reduce risks, that improve efficiency and that deliver valuable data to drive competitiveness. The icing on the cake is that very often, these tools come at little additional cost because they leverage new value from existing investments.