Modern, video-based solutions can capture footfalls, employ facial recognition and analyse movement and human behaviour.
For several decades, video was simply a device to record actions and movements – mainly used in entertainment and security applications. Till the nineties, video meant analog – good only for viewing. Tape storage ensured that archival and retrieval remained cumbersome.
It is only post-2000, that the now ubiquitous digital cameras took off. The ability to store (and manipulate) video data digitally laid the foundation for video analytics. There have been simultaneous and significant advances in chip technology, cloud storage, machine learning, artificial intelligence, and fast and affordable bandwidth. It is now possible to move, process and analyse video content in real-time.
So What is Video Analytics?
Video analytics essentially enables the computer to understand video data. We can compare with archives, analyse metadata in multiple ways, build algorithms to feed AI systems… and more! There are numerous analytical tools that allow software to detect motion, shapes, objects, colours; recognise faces and pinpoint locations. And it can now happen in real-time. But the x-factor is that the machine is learning all the time, and that is what drives the next video revolution.
Where Can Video Analytics be Used?
Traditionally, one of the major applications of video has been surveillance. And this segment has naturally seen some of the fastest adoptions of the new technologies.
Analysis of movement or change in video images allows generation of real-time alerts in case of illegal entry (perimeter threat detection), or when customers enter a shop, or when a worker strays into an unsafe area. Similarly object detection is used to spot weapons, unaccompanied bags at airports or in trains and buses, out-of-place objects in factories or construction sites. Similar systems can detect fire and smoke. Cameras have audio capability as well, and unexpected sounds (e.g. screams) can trigger alerts.
Motion tracking technology is being used in the gaming and film industry to mimic real movement and enhance animation quality.
Not Just About Surveillance!
But security is only the tip of the iceberg, as new and innovative use-cases are being imagined and created at a furious pace. Here are just a few examples of what’s already being done:
Probably the most controversial of video analytics tools is face recognition. Reports suggest that the Chinese government is now able to recognise and track its billion-plus citizens as they move around. Even as many fear that such technologies will be misused, the reality is that the genie is out of the box. Today, even your phone comes with a face-recognition feature. The security benefits are obvious, and such systems are used in high-security establishments, airports, border control, etc. It’s not too far-fetched to think that offices or factories will increasingly use such technologies for security and attendance. And many governments will try to emulate the Chinese.
Video analysis can now capture traffic violations and recognise number plates simultaneously. In many cities, such solutions are enabling automated tickets, tolls, parking, and entry to restricted areas.
Modern video-based solutions can record and analyse footfalls. These systems are used in retail outlets, airports, stadiums, museums, and all kinds of public events. Apart from the obvious crowd-control applications, retailers can better understand shopper footfalls and conversion ratios by time of day, or day of week, or during special occasions. Such technologies are being used to enforce social distancing and queue management, preventing crowding at malls.
Taking this a step further, in stores; algorithms or “heat maps” generate insights on where customers went, how they navigated, what they looked at, for how long, what they handled, checkout time, etc. Apart from improving product placement and store layout, with personalisation algorithms, retailers can significantly enhance customer experience and sales conversions. Similar visitor movement data and analytics are creating opportunities in a wide variety of applications.
In factories, apart from the obvious security and safety applications, video analytics can help detect flaws in products and machine operations – in real-time! Motion tracking of goods or equipment in warehouses, ports and factories can improve traffic flows, enhance productivity and reduce inventory costs.
Film-makers and gaming companies have long used motion tracking technologies that learn from real people movements and enhance animation quality. But now, new and exciting applications are exploding.
In cricket matches, ball-tracking technology is used to help umpires predict the likely trajectory of the ball in leg-before-wicket decisions. Golfers use videos to analyse their own actions and perfect their swings. In many competitive sports, video analytics is used by coaches to analyse their own teams as well as that of rivals.
Elderly people in old age homes or living alone are being monitored using video and abnormal behaviour triggers alerts to doctors or healthcare workers.
A variety of video analytics tools are what makes it possible to have driverless cars. Signals from multiple cameras are analysed in real-time, to make sure the vehicle doesn’t break a light or hit a person or another car, or fall into a ditch.
And it doesn’t end there – Computers are even analysing human behaviour from videos! Facial expressions and body postures are used to assess emotions and allied behavioural patterns. Some algorithms can be applied to groups of people, encouraging use in law enforcement and crowd control. Employers are using video analytics to assess candidates and their responses. Were they lying? Or uncomfortable while answering a particular question? Similar technologies are used by marketers and market researchers. Smartphone apps have been developed that can help detect autism spectrum disorder (ASD) in children, by analysing facial expressions and eye movements.
And This Is Only the Beginning!
As video analytics technologies evolve, we will see applications that we cannot imagine today. The COVID crisis has substantially accelerated this evolution as video communication is now the norm – for meetings, education, marketing, recruitment, conferences… and much more.
At the risk of sounding repetitive – this is only the beginning. Video analytics is fundamentally disruptive, and will alter our lives in ways we cannot even think of today.
Arun Jethmalani, Founder & Managing Director of ValueNotes, has a 35-year working career spanning IT services, management consulting, market and equity research, competitive intelligence and web publishing. He has been extensively published in leading newspapers and journals on equity investing, strategic intelligence and technology; and is a regular speaker at global conferences on research & intelligence, technology, finance and tech services.