Exclusive: Why AI video analytics are needed more than ever
Share this content
Today’s AI video can build resilience but only if users understand what AI really is, writes Jamie Barnfield, Senior Sales Director, IDIS Europe.
Deep learning AI-enabled video is one of the best tools that organisations can now deploy for improving their operational efficiency and future resilience.
But not all video analytics are designed to the same standard or deliver equal value and the terms ‘intelligent’ and ‘Artificial Intelligence (AI)’ have sometimes been applied very loosely.
It’s important for users to understand the differences, now more than ever. The global economy is entering a turbulent phase. Yes, there may be greater optimism around the expected recovery but until vaccinations are delivered worldwide the risks and uncertainties will remain high.
A continuing threat
In February, even as vaccination programs gathered pace in richer countries, a report from scientists in California showed why vigilance will need to be maintained: a first ‘recombination’ event was detected, whereby two mutated variants of the SARS-CoV-2 coronavirus had combined to form an entirely new hybrid. This process is different from the mutation routes we’ve seen so far, according to a report in New Scientist – and it raises the possibility of evolutionary shortcuts being taken as more infectious variants combine with more resistant ones.
That pessimistic outcome is only one possibility of course, but the recombination mechanism is well understood and it has been expected to occur – the risk is higher as the virus is so widespread – so this news serves as a warning.
Scientists simply can’t know whether SARS-CoV-2 will evolve to become more dangerous or less. But there is general agreement that there will be future infection waves and other pandemics.
So, as they look to invest for recovery and growth, organisations should also seek practical and affordable ways to improve their preparedness and resilience.
Upgrading to true AI video does both and it fits the ‘build back better’ agenda.
Better use of video data
It’s well understood now that, to run sites safely and efficiently, it’s necessary to have good data – for example, on the movements of people, on area occupancy and density levels and on compliance with hygiene control measures such as mask-wearing and social distancing. The latest generation AI video analytics can do all these things and, in settings such as retail, they can be integrated with automated displays at entrances showing wait times or can trigger recorded announcements reminding people to follow hygiene guidance.
As well as building resilience against infection risks, the latest AI video is also better at tackling those ‘traditional’ challenges of crime risk reduction thanks to enhanced accuracy for tasks such as loitering detection, intrusion detection, object tracking etc. For control room staff that means improved situational awareness and greater speed identifying, verifying and responding to threats – in other words, a proactive approach to safety and security rather than waiting for incidents to play out.
The best AI-enabled video solutions provide compelling tools for the most important threats organisations now face, they deliver more data in more granular detail than ever before and also make that data more manageable.
From ‘blobs’ to neural networks
So, what do we mean by true AI deep learning analytics?
Many of the early iterations of video analytics rely on Binary Large OBject (BLOB) technology. This is still found on most modern IP cameras, which is why they are commonly referred to as “blob type” analytics. These are formulated to detect an event, such as a virtual line cross; they detect and track objects as ‘motion blobs’ and distinguish them from smaller binary objects. For many applications these are still useful. But video analytics capability has moved considerably beyond this.
Today, security departments can take advantage of deep learning that leverages neural networks made up of multiple layers of algorithms and advanced processing. This is now driving what is widely accepted to mean true intelligent video analytics. Deep learning engines are ‘trained’ using vast datasets of images and video footage of people, objects and vehicles. They can ‘look for’ size, shape, speed and directional information and they continue to learn while in use. To an extent, deep learning replicates the way neurons work in the brain: it can analyse and prioritise input from video data to decide which inputs are of value and it will notify security operatives accordingly.
Deep learning’s real value comes from being able to detect suspicious activity or unusual events and eliminate those smaller binary objects that are just “noise” and apply rules that meet with specific applications and operational requirements. In addition, deep learning should enable users to use metadata to search multiple camera streams to find the most accurate matches for persons or vehicles of interest within minutes.
A need for caution
But again, some caution is needed. Deep learning video offerings can still disappoint, generally as a result of having been launched too early, before engines were fully trained and able to recognise objects reliably and accurately. Systems integrators and their customers need to exercise caution regarding claims and jargon such as the difference between machine and deep learning. They need to understand for how long – and using how much data – neural networks and algorithms have been trained. Both make a difference. And they also need to be clear about which offerings and functions will genuinely add value and help them to increase productivity, provide useful business intelligence and deliver ROI, long term.
For example, traditional blob type analytics cameras are prone to being triggered by environmental factors, such as heavy rain, snow, or moving foliage and struggle to distinguish a human presence, which may present a threat, from harmless animal activity. For users, this is usually the main driver for adopting AI-assisted notifications because false positive alarms can result in time being wasted investigating the cause of alarms and the larger the site, or more overstretched the system operator, the worse that problem can be.
Unlike human brains deep learning engines don’t get tired. They can constantly monitor multiple camera streams in search of suspicious behaviour, maintaining performance levels even in the busiest scenes such as retail malls, logistics centres, higher education settings and outdoor spaces.
All of this means that systems integrators need to test accuracy not just on the performance of analytics rules but metadata searching rather than relying on claims or vendor-led demonstrations that often take place in controlled environments. The better the accuracy and speed the more productivity gains users will realise and more importantly it avoids disappointing customers and means they choose a future proof solution upon to tackle future security and other operational challenges.
Better protection now – and resilience for tomorrow
AI-assisted notifications free-up operators from having to constantly monitor multiple camera streams and video walls, so can transform the work of security teams. Today’s truly smart video technology can allow security provision to be better focused, with officers being re-deployed to more important tasks that add greater value to their roles. And strategically, heads of security can interpret and use accurate real-time and historical data to drive more informed decision-making, to better mitigate risk and provide actionable intelligence across their enterprises.
AI video also offers value beyond security into other departments, such as facilities management, helping to optimise workspaces and buildings or benefiting heads of marketing in retail and hospitality with better understanding of customer behaviour.
If COVID-19 has taught us anything it is that businesses need to be able to pivot and adapt quickly. Flexible and scalable AI solutions are a compelling choice, particularly if they offer backward and forward compatibility, leverage existing investments in cameras that can be used for security and safety as well as business intelligence purposes, as well as working with VMS that allows integration with third party systems and devices, plus correlation with other data.
This article was originally published in the March 2021 edition of International Security Journal. Pick up your FREE digital copy here