Who is the man in the mask?
As the world starts to get back to business, John O’Brien explores the role that digital signage will play in keeping us safe. But is it all a bit Orwellian?
Times of great disturbance create opportunity for quick thinking and rapid innovation. With our current societal upheavals and consequent physical restrictions, a plethora of technology solutions are emerging to keep us safe in public.
From the start of this tumultuous year, conferencing and streaming services have prospered (along with more mundane sectors like cleaning and freight logistics). Now, as we all gingerly venture back out and about, clever targeted signage is increasingly important to maintain appropriate physical distancing in communal areas.
What sign are you?
It would be fair to say that public space advertising and information sharing has undergone a significant revolution since the advent of digital content provision and management. Static hoardings still have their place but the flexibility and availability of digital signage are hastening their demise.
Retail advertising, information, promotions, wayfi nding instruction and now public health communications are all rich fodder for intelligent notice boards.
Media delivery to these screens has evolved from secondary disc-based players to USB drive updates to centralised streaming and Content Management Systems (CMS). Old timers might remember driving around to their installations every month to insert the content disc that had been freighted from overseas!
Whichever way the material is served, screens display in only one direction – to the eyeballs of the intended recipient. Add some feedback from the viewer or environment and the whole relationship becomes very dynamic and much more powerful.
Inputs & Outputs
A simple input device, like an occupancy sensor, can let the CMS or related system know that one or multiple persons are within range. From that basic knowledge, the signage designer / administrator can tailor the display content.
Many types of inputs and their associated data are useful here. Thermal sensors, proximity sensors and cameras can feed information into the nexus. Data from building management systems, government authorities and other sources can also aid the automated decision-making process in what to show at any given time.
Practical integration of inputs and outputs in the public display space has been under way for a while in the retail sector. Pre-restrictions, Leeds Playhouse in the UK incorporated their fire alarm system into newly upgraded signage. In the event of an emergency, all screens add simple visual instructions for safe evacuation of staff and customers.
Retail wise, In-store foot traffic analytics already use heat maps to show high and low flow areas. Demographic analysis of shoppers also helps increase the effectiveness of targeted content and display positioning. Until recently, this had seen a slow and steady roll-out.
Only last year, this sort of tech was assisting retail and venue enhancement. Within a few short months, it has transitioned to encouraging compliance and keeping us all safe.
Area occupancy, space utilisation, traffic flows, and count based rules are all key parameters in analysing store trends but are equally applicable in reducing crowding and maintaining appropriate distancing. Pandemic restrictions have rapidly accelerated the development and deployment of technical solutions that facilitate this interactive engagement with the end user without having to touch anything.
Many IP cameras are already smart enough to detect motion, temperature, line crossing, distancing between people and tracking people or objects throughout their view. This information can then be relayed to everything from local screens to remote command and control centres.
At a simple level, products like LG’s Wellness Kiosks are standalone units with embedded thermal scanners or InfraRed cameras that can trigger warning content based on a high temperature reading.
A more elaborate example is Phillips PeopleCount, a collaboration with camera and signage providers. Using a Bosch camera that can count people interfaced directly to a display with onboard CMS, the setup can show pre-determined customer information or warnings based on spatial distancing. This works for a single entrance.
Add a network connection to the display and the real power is unleashed with the CMS and reporting software. Phillips also have a multi-entrance solution with HikVision and NowSignage that adds centralised tallying and total store (or sector) head counts from cameras at each entrance. This count can trigger the display to show “Store is currently full”, “You are X in the queue.” and the like.
Recent reports from industry suppliers are of a heavy demand for network ready signage product. Fittingly, retail and commercial outlets are desperate to get foot traffic through the door while equally desperate to do their bit and not be a vector for disease spread. Their motivation may be either commercial or altruistic but the outcome is the same – smart signage is proliferating and it is not standalone any more.
The eye in the sky
With the development of IP camera technology, CCTV has become particularly influential. When paired with information outlets, this becomes a powerful method of communication. When real time personal recognition is added to the mix, it becomes a potent soup indeed.
China’s large-scale civilian surveillance reportedly uses roughly 200 million CCTV cameras to cover 1.8 billion people and there is some very sophisticated technology behind this arrangement. But the UK was also an early leader in public camera coverage and now has 5-6 million eyes trained on public and private spaces. In Europe, GDPR has tightened up how the data gathered from these devices can be used but they are still omnipresent.
Australians were once a pretty suspicious lot when it came to state intervention in their lives. CCTV had previously been met with distrust but Sydney now has over 300,000 closed circuit cameras in operation. Federal legislation in the area of biometric data protection is currently lacking and there is a push to allow deeper scrutiny of the populace by our state spooks.
Individual privacy is always a trade-off with collective good and responses to recent government emergency edicts has been mixed but generally well accepted. Given that we’ve already sold our souls to the social media giants, how concerned should we be by Canberra?
We know who you are
Which brings us to the next level up of gadgetry – individual recognition and identification. China has quite extensive and elaborate methods in place for citizen tracking. Much of the power in their machinery comes from real time facial recognition. And if they can’t see your face, they can just as easily identify you through your gait.
This gear is far from limited to the middle kingdom though. Law enforcement agencies and militaries worldwide are investing enormous amounts into tracking and pinpointing apparatus. Some of it is in the public sphere, some of it is covert.
To help gain some statistical evidence, France recently rolled out changes to CCTVs installed in Metro buses, adding small CPUs that process the video content in real time, packaging it up for download. It has just enough smarts to identify mask or no mask and the aggregated information is automatically sent to the relevant authorities at the end of each day. Pretty harmless and pretty helpful.
How does this work?
What those little computers on ‘les bus’ are doing is Video Content Analysis (VCA) – automatically looking for faces with masks or not, all without human intervention. This instance happens in-camera (at-the-edge) but the really powerful setups have dedicated centralised processing hardware and software.
As an example, I looked in detail at a proprietary VCA package targeted primarily at US law enforcement agencies. BriefCam has many capabilities beyond post operational video review. It can single out an individual object in a frame and trigger alerts based on that or any other selected parameter, and, powerfully, it can do this across multiple camera sites in real time.
This is but one of many real time VCA solutions on the open market. I can imagine several national governments already have in place at least similar capacity. While there are multiple commercial entities operating in this space, there are also many open source efforts going on. I’ve found numerous examples of home-cooked VCA coding that can analyse video footage looking for facemasks. Where they will get used it’s hard to predict but it does illustrate how possible this is to do with limited resources. With more formidable wherewithal, the potentials are huge.
These analytics methods use plenty of preset algorithms but also increasingly effective AI and neural learning networks to pull apart all the little bits and find the desired needle in the haystack. Some of the analytical processes are post-event (after the fact) but so much now happens in real time.
Given that the pace of technological progress seems unlikely to slow in this field, I’m both intrigued and scared by the possibilities.
Where’s it headed?
A pessimist might see Orwellian implications in the convergence of tech and surveillance but our commercial and public sectors are jumping right in anyway. Crystal balls are notoriously unreliable though … maybe AI can see the future?
Regardless, human beings are both resilient and resourceful and we will find a way to make the best of the current bad situation. It is pretty clear that interlinked public signage and tracking will become more prevalent in the future, as will the independent machine learning behind it. How pervasive that becomes is up to where each of our societies and their businesses allow it.
As we travel this unfamiliar road, a canny integrator might find good opportunities with retail and public space managers. Make the most of the upheaval and use your tech and business nous to help keep us all safe.