How smarter AI™-powered cameras can mitigate the spread of Wuhan Novel Coronavirus (COVID-19), and what we’ve learned from the SARS outbreak 17 years prior.

As the World Health Organization (WHO) declares the coronavirus (COVID-19) outbreak a ‘global health emergency,’ we examine the thermal camera sensor setups that dealt with SARS in 2003, to see what’s changed 17 years on, in dealing with the Wuhan Virus outbreak. Finally, we explore the missed opportunities to deploy AI-driven smart camera solutions on border camera setups that are over a decade and a half old, which could improve the accuracy, efficiency and scalability of containment efforts.

How did thermal camera sensor checkpoints look like when SARS came around in 2013?

In 2003, the world grappled with severe acute respiratory syndrome—officially designated as SARS coronavirus (SARS-CoV)—which culminated in 8,096 cases reported in 29 countries, and 774 deaths. At the time, governments the world over responded by deploying thermal imaging sensors at border checkpoints (airports, seaports, border crossings) to check for one of the key symptoms of SARS; a fever.

Countries like Singapore even innovated by deploying military-grade technologies to their screening checkpoints, enabling more precise thermal imaging than the commercially-available radiometric thermal scanners of the time. 

Sensor hardware aside, the setup was really straightforward. A camera sensor is connected to a monitor and employee(s) with accompanying personnel monitoring the terminal with thermal sensors and a screen that displays the feed in real-time. 

At the time, this was a revolutionary step in “contactless” monitoring as it did away with the time-consuming personal temperature readers that required employees conducting screening to come in close contact with individuals. 

Measures like thermal screening (one of the key strategies employed to identify affected individuals crossing borders) were deemed effective in the mantra held by health officials; to identify, quarantine, and treat individuals with SARS to avoid a global pandemic. It proved effective at the time, as a fever 100.5 F (38 C) being a hallmark symptom of SARS was easily identified by thermal scanners. By July of 2003, 6 months after the first case of SARS, the World Health Organization (WHO) declared the virus contained.

Between then and now, health officials also had to deal with the Middle East respiratory syndrome coronavirus (MERS-CoV) in 2012. With systemic fever also being an accompanying symptom of the virus, the same measures were once more deployed and proved to be an effective tactic for mitigating cross-border spread. However, viruses are ever-evolving, and in 2020 health officials who direct screening protocols are faced with a new predicament: what if temperature screening alone is not enough?

 

Have thermal camera sensors become smarter in 2020, to deal with the Wuhan-originated novel coronavirus (COVID-19)?

Fast forward to 2020, and we have a new coronavirus that has authorities scrambling to detect and mitigate. Known by a few names: coronavirus, Wuhan flu, 2019 novel coronavirus, 2019-nCoV (COVID-19), there is, as of writing, 9,822 known and confirmed cases across 23 countries, with new cases still accelerating, and new countries being placed on a global infection list on a daily basis.

Once more, the world turned to a tried and tested setup. A thermal imaging camera hooked up to a monitoring station, that needed to be physically monitored on-premise. As time would tell, applying the same inflexible solution to a crisis a decade and a half old, might not be sufficient this time around. There has been little innovation on the front of thermal imaging stations. 

So we get to the question: have thermal sensors become better or smarter? In some ways, it has. Changi Airport has deployed more mobile temperature checkpoints, that are no longer in fixed positions. This allows the airport to strategically place terminals at high-risk inbound flights, and allow for more granular scrutiny of each arriving passenger from different points, instead of watching every single passenger through a single checkpoint. 

Additionally, it seems like there is some degree of software functionality built into the systems deployed right now. On the functionality tab in the image above, there are visual alarms, audio alarms and area effects that help bring greater attention to key individuals. It looks designed to help the watchful eyes of the personnel attending to the thermal stations. There is also an indicator of system health that could signal when terminals need replacement or maintenance, as well as differing levels of view based on the user type. 

These new functions are no doubt positive steps in enhancing features of such terminals, but is it enough? As the ever-increasing numbers of cross-border cases have indicated, they aren’t. 

 

Missed opportunities in efficiency, accuracy, and safety, due to the lack of smarter camera sensors.

The rudimentary setup of health checkpoints seems to have evolved little beyond enhancing usability for the individual monitoring the screens, since 2003. This is a missed opportunity at best and outdated practices at worst. There are technology solutions available in the market that are automated, AI-powered and combine multiple inputs of sensor data, that could go beyond just checking for temperatures.

Let’s explore how.

 

Problem #1 with thermal scanning checkpoints today:
They need to be manned posing a staffing challenge to keep these terminals running 24/7. 

Smarter cameras of today tend to be wirelessly connected to the cloud, enabling remote monitoring of checkpoints. With smarter cameras on the thermal checkpoint setups that we have today, we can do away with the need to have each station physically monitored on-site. It can be done remotely, therefore consolidating manpower needs to enable comprehensive thermal checkpoints. 

This is a real-world problem plaguing airports like Changi Airport in Singapore, one of the countries with the most infected cases outside of China (13 as of 31 January 2020). We came across a Facebook post where Changi Airport needed “500 Healthcare Assistants urgently.” This shows the challenges in keeping the various thermal checkpoints manned in airports that don’t sleep. According to a corresponding news report in a local Singaporean newspaper—only half of these posts have been filled. There is no use in having hardware that cannot function when they are unmanned. 

Utilizing smart, AI-driven cameras with cloud/edge inference could perform the same tasks, doing away with staffing needs by automating the job at hand, 24/7, on-demand. They could be set to automatically send alerts to people on-ground when necessary (such as when a high fever is detected). This could also reduce the chance of human error, and avoid detection lapses resulting from inattention. By significantly easing the operational burden of having to train and staff the terminals, deployments can be done more efficiently and at a much larger scale. 

Finally, this would also mean that healthy individuals need always be in close contact or proximity with the high-risk individuals that they need to monitor. Connected smart camera sensors can be viewed remotely, from anywhere, such as a central command center, from where efforts to identify and act on individuals with symptoms, can be coordinated.

A central command center powered by smart connected cameras, from where monitoring efforts are consolidated (and assisted by automated, AI functionalities) would drastically reduce staffing requirements to comprehensive deployments required by such crises, and allow operations to scale (upwards or downwards) as needed, and not be held back by a shortage of personnel.

 

Problem #2 with thermal scanning checkpoints today:
Fixed, often single/limited functionality of thermal checkpoints that function as nothing more than visual thermometers. 

Fever is a symptom of the Wuhan-originated coronavirus. Many countries have installed comprehensive thermal checkpoints. Most cases of the virus are imported from infected individuals traveling into a country from China. If these points held firm, checkpoints should serve as an airtight filter, but it doesn’t. 

In fact, the top 5 countries with the highest number of recorded confirmed cases are Japan (14), Thailand (14), Singapore (13), Hong Kong (12) and Australia (9), which each have comprehensive thermal checkpoints in their airport. There could be a few factors for this. 

One could be a point we brought up in ‘Problem #1.’ Insufficient manpower could lead to loopholes in comprehensive setups. There might still be individuals who slip past detection efforts because of this. Human error should also be considered, either from carelessness or inattention. 

Another likely culprit is the evolving nature of the virus over its predecessors MERS-CoV and SARS. It has been reported that there are more “varied” symptoms of the infected, compared to their predecessors, which means a fever is not always a present symptom despite them having pneumonia. A solution could be the utilization of sensor fusion, where thermal cameras are used in tandem with regular cameras with added AI functionalities. Such cameras can be trained to also detect behavioral ‘events’ or ‘instances’ such as someone sneezing or coughing—other symptoms of the virus. 

Another functionality of smarter cameras is the ability to identify and log individual faces that display any number of symptoms it is trained to recognize. Through advanced video telematics and sensor fusion, enhanced by AI inference, efforts to more accurately identify individuals could be made more foolproof. 

 

Problem #3 with thermal scanning checkpoints today:
Without direct coordination with city-level databases, efforts remain isolated if it does not relay and coordinate critical information captured from non-smart devices. 

Detection efforts today are managed by separate entities like checkpoints (land, air, and sea) that may not have information being logged to be accessible to other parallel efforts at mitigating the spread of the Wuhan coronavirus. 

This is especially true if the devices in use are non-smart stand-alone devices that are not connected to a central database, and not capturing vital information about the movement of suspected individuals. Smart connected cameras have the ability to trigger specific event recording when an “event” has been recognized (such as a person sneezing, coughing, or running a fever). With event recording, authorities need not pour through traditional CCTV footage to back-trace individuals.

There are also opportunities to synchronize the data such that devices essentially “talk” to each other through the cloud. If we add to the fact that most cities today have comprehensive surveillance networks, if systems that stand as silos today communicated information as a group, there is the benefit of shared data from which authorities can act on. This could theoretically make identification of individuals with any number of suspected symptoms be easily identified and tracked. Going beyond this, smart connected cities can also identify who came into contact with infected individuals, making on-ground efforts like contact tracing of afflicted people at risk, much quicker. 

 

17-year-old solutions that are neither smart nor connected, deployed in a highly-connected world is insufficient in dealing with the Wuhan Coronavirus (COVID-19).

The varying opportunities to employ smart connected cameras during times of crisis could help city officials or entire countries, better coordinate efforts in minimizing the impact of global emergencies like the Wuhan coronavirus. 

The technologies that enable advanced AI-driven functionalities that health officials would have deemed science fiction when SARS first broke in 2003, is already available today with platforms like AnyConnect. It’s a shame that 17 years on they have not fully utilized it, and global emergencies like this signal that it is perhaps high time that they do.

Updates:
Article updated 14/02/2020 to reflect the official designated name for the Coronavirus: COVID-19.