Masking Accuracy: The New Facial Recognition Challenge for Airports
Mandatory mask-wearing is causing facial recognition problems, impacting passenger flow. How can airports keep travellers moving at safe distances, without reverting to manually counting heads?
Masks are considered one of our best defences against the spread of COVID-19, including during air travel. People are required to wear masks from entering the departure area until exiting the arrival. However, this has given rise to another problem for airports — how to recognise faces when they’re covered up.
Facial recognition at processing points has become widespread in recent years. In the US, border protection agency CBP hopes to biometrically screen 97% of all arriving international travellers by 2023. With the pandemic accelerating the move to contactless processing, the technology’s potential to support seamless access through the terminal via a digital identity is promising.
However, the shift to a fully masked population has led to some challenges. This is especially notable where the technology relies on tracking moving faces to manage queues and understand flow. As reported by the Washington Post, a study by the US National Institute of Standards and Technology (NIST), shows that masks covering the nose and mouth can cause an error rate of between 5 and 50% in many facial recognition algorithms. The more of the nose that’s covered, the harder it is to identify the face, especially when the person is moving.
How to smooth passenger flow during pandemic times
As a result, many airports are measuring queue wait-time manually. Unfortunately, this puts significant stress on front-line staff and already strained budgets. Is there another way to accurately manage passenger flow through checkpoints, and around the terminal, without relying solely on facial recognition?
Customer analytics platforms that use a combination of location analysis technologies are bridging this urgently needed information flow gap. By meshing data from multiple IoT sensors and airport systems, operators not only have an option to accurately and cost-effectively analyse dwell times but also predict behaviour and crowding. All while protecting passenger privacy.
There are a number of reasons why the tech combination works for our changing travel situation.
1. Greater traveller confidence
With many unknown factors surrounding air travel right now, travellers want to know how quickly they will move through queues to a safer space. One of the best ways they can feel safer, and in control, is through real-time information-sharing.
By capturing accurate real-world data from various physical environments (inside, outside, high or low ceilings, narrow spaces, different lighting conditions), airports can build the most detailed picture of queue length. In addition, they can evaluate passenger movement and spacing in real time — from drop-off zones right through to checkpoints and overflow areas.
With this information, airports can communicate accurate wait-times to people in queues and to those who are about to join or guide on how to avoid crowded areas. With this greater sense of control, travellers feel reduced stress levels, and airports build greater trust in their safety measures.
2. Faster decision-making
As mandates change — introducing COVID-19 testing, health screening or physical distancing — airports constantly need to reconfigure available spaces.
Connecting location and movement analysis with flight information, allows operators to understand much more about how passengers arrive and how queues form. They can identify when they enter, where they gather and for how long. They can understand the breakdown of each line by flight and forecast the impact on the queue of any changes to the flight schedule. With this insight, they can then make fast, accurate decisions to improve flow, reduce crowding, and adequately staff each area.
Simulation tools can also use this data to test the impact that changes, like COVID-19 requirements, have on different processes. This allows airports to review possible scenarios, such as the addition of health check stations, changes to flow around check-in kiosks, or additional lane openings before they are implemented.
3. A low-cost, airport-wide alternative
Queues and crowds can form anywhere, not just at check-in or security. Gates, lounges and baggage carousels are also potential risk areas where people may not be able to physically distance.
Digital-first airports, like Amsterdam’s Schiphol and Iceland’s Keflavik, have adopted a terminal-wide approach that analyses behaviour data from kerb-to-flight. As a result, they are able to quickly and affordably scale to gain detailed occupancy and dwell insights and forecasts around the terminal. Critically, this insight also enables airports to understand how one process, like passenger processing, can influence others downstream, like on-time departures.
This, in turn, enables smarter decision-making to implement efficiency improvements, and support distancing — from redirecting flow and call-to-gate times, to spreading gate and baggage belt allocations.
Trusting tech to keep passengers safe
While the use of biometrics has many benefits for contactless passenger processing, airports need to urgently find a reliable way to get masked travellers quickly and safely through the terminal.
That’s where remote sensing can complement a biometric strategy. It’s quick to install so operators can accurately measure queue times, density and movement within weeks — whether or not masks are worn
At these times, airports need to trust the tech. For strained operators and beleaguered passengers, that confidence will make all the difference.