FEATURE: ARTIFICIAL INTELLIGENCE
they refer to different algorithms. ML, for
example, is a branch of AI which utilises the
concept that systems can learn from data,
identifying patterns and improving processes
or making decisions, with little to no human
intervention. Whereas DL is a sub-category
of AI that uses algorithms inspired by the
structure and function of the brain, called
artificial neural networks.
But how does this apply to security and,
specifically, video surveillance?
AI-powered accuracy
AI has the potential to bolster core
functions of security, most prominently
CCTV and video surveillance. CCTV is by
no means a new technology – but, when
combined with analytics-based AI, it can
bring new improvements to the security
function of a business.
When analytics are incorporated within
video surveillance, CCTV is transformed from
a passive tool that simply records events,
to an active one which can help teams
quickly identify suspicious events. That
being said, analytics-powered detection
has been known to erroneously pick up
on natural events and mistake them for
potential threats. These false alarms are one
of the biggest challenges faced by security
Kevin Waterhouse, Managing Director
at VCA
professionals when it comes to analytics in
video surveillance – they unnecessarily call
upon staff members’ attention when they
could be working on more pressing matters,
causing the waste of time and resources.
So how can these systems’ accuracy be
improved? That’s where AI comes in.
Deep Learning, a subset of AI, can be
deployed to ‘teach’ the system how to
distinguish humans and vehicles from, for
instance, a wind-blown tree branch – thus
differentiating between a false alarm and
a real threat. By calibrating the system
to identify different moving objects, the
machine can assess which events to flag
to security teams, who can then determine
how to address them. This way, only events
that can truly pose a threat to the business
create an alert, and staff members’ time and
efforts can be focused on real security issues.
A timely help
The COVID-19 outbreak has undoubtedly
had a considerable impact on the current
business landscape across sectors, and
security is no exception. With more than
a fifth of the world’s population under
lockdown, many buildings are left empty
while employees work remotely. It’s clear
that security risks are heightened. To make
matters more complicated, as workers are
furloughed or encouraged to stay at home
to facilitate social distancing regulations,
security teams are depleted and business
premises have become even easier targets for
those looking to exploit these vulnerabilities.
It’s never been more urgent to enhance
video surveillance systems. Firstly, analyticspowered
surveillance solutions can take over
the onerous job of constantly scanning live
footage to spot unusual behaviour and alert
the few remaining staff of potential intruders.
With short-staffed security teams, the help of
analytics-based threat detection that triggers
an alert and determines when action is
required is simply invaluable.
56 INTELLIGENTCIO www.intelligentcio.com