Deployment, Control, and Evolution of Biometric Services
Despite some complications, such as regional coronavirus variants and the ongoing conflict in Ukraine, the world economy continues along an uncertain path toward recovery, in the wake of COVID-19. All sectors are coming back to life in some capacity - including air travel.
Speaking recently at the US Travel Association's 2022 State of the Travel Industry Address, Roger Dow expresses a degree of optimism for the industry - particularly in the business travel sector.
Even though business and international travel are not projected to reach 2019 levels until 2024, Dow believes that the sector will bounce back much more quickly than economists predict. This recovery will hinge on a policy agenda and strategy that includes “strengthening the workforce and restoring jobs, facilitating seamless and secure travel, highlighting the future of travel mobility, prioritising sustainability; and focusing on diversity, equity and inclusion.”
Economic recovery must look to business travel and physical interactions to accelerate its progress. Facilitating such travel - and enabling the seamless experiences that contemporary travellers now demand - will involve technologies like trusted traveller programmes and biometric services.
The Scope and Role of Biometric Services
Biometric services encompass a range of identity verification and authentication techniques, based on the unique physical and/or behavioural characteristics of an individual. Various modalities exist for implementing biometric services. They include finger vein, palm vein, fingerprint, facial recognition, lips, iris, and retina scanning.
Europe’s Policy Department for Citizens’ Rights and Constitutional Affairs defines ‘biometric techniques’ as “any technology or operation that relies on specific technical processing of data relating to physical, physiological or behavioural aspects of the human body” (including when in motion).
Beyond the traditional biometric techniques such as fingerprint or facial recognition, biometric techniques also include areas such as signature dynamics, gesture dynamics, and the analysis of keystrokes or mouse movements. Recent technological advances for biometric services include the rollout of improved sensors, which enable the capture of entirely new types of biometric signals, such as heartbeats and brain waves, and the development of brain-computing interfaces (BCI).
Biometric services and technology are gaining traction and scope in several areas, including law enforcement, border control, and traveller management. Here at Vision-Box, all biometric products rely on biometric devices (hardware devices with integrated biometric engines) to be able to capture and validate a person.
While modern systems are designed for consistency and accuracy, relying on a single characteristic as the basis for identification can be risky. Using multimodal biometrics to authenticate individuals from a variety of distinct identifying traits improves both accuracy and security.
Our touchpoints work with two security vectors. One of them is the ID document that a person presents, versus its live identification. We incorporate in our systems advanced state-of-the-art technologies, both for fingerprint and facial recognition.
Why Biometric Services are Expanding in Scope
For travellers, biometric facial recognition and its related technologies help in facilitating frictionless travel experiences, in which people can spend less time in queues and more in airport services and amenities. Biometric identification services effectively eliminate the need for additional forms of physical identification, such as paper-based passports, boarding passes, and tickets. Moreover, reducing the number of physical contact points by using biometric facial recognition enables ports, airlines and other types of carriers, as well as border forces to reduce the possible spread of COVID-19 and other pathogens in the airport environment.
According to the 2021 Global Passenger Survey conducted by the International Air Transport Association (IATA), passengers are prepared to use identification based on biometric facial recognition, if it will speed up and smooth out their travel processes. 73% of passengers say that they are willing to share their biometric data to improve airport processes.
However, there are caveats. In cases where biometric services fail to accurately identify a traveller, the individual may be denied rights and privileges that they are entitled to enjoy. In addition, biometric identification algorithms must perform in the same way regardless of the population’s age, gender, or ethnic group - that is, without algorithmic bias.
The onus is therefore on ports, carriers, and border forces to get their implementation of biometric services right.
Getting it Right
A biometric-based algorithm is the set of rules governing Machine Learning (ML) and Artificial Intelligence (AI) which provide step-by-step instructions on how biometric systems should complete specific tasks related to biometric recognition. Algorithmic bias can become an issue if algorithms contain rules based on preconceived ideas about a certain person or group or lack sufficient data to adequately represent an entire population or target audience. This might for example be based on ethnicity, language, or skin colour.
To increase the confidence level of an algorithm, it is imperative to segment datasets both geographically and demographically, with faces of different ages and genders, to train the algorithm. This enables system designers to increase their level of security when performing the matching test, i.e., ensuring that the face scanned is that person.
At Vision-Box, we invest in the ongoing work of the algorithm to ensure that it is ethnicity, age, and gender agnostic. So, the elimination of algorithmic bias becomes an integral part of the biometric lifecycle management process.
We develop our algorithm in-house. We believe it’s crucial to have all the information regarding what we sell on our side. This put us in the position of giving our customers the best experience and the best product.
When we buy technology from a third party, we never know what's “inside” - and we can't tamper with or manipulate the algorithm. This brings executive and business continuity limitations. As part of the biometric lifecycle management, we must be responsible for the continuity of the components that make up our entire solution. If we buy technology from a third party and there is a situation beyond our control, we may lose access to the technology we are reselling.
By integrating our algorithm and delivering a product that is 100% ours, and built by us, we are able to provide better client support. We can check and implement updates and changes - and provide those changes much faster, as we are not relying on a third-party company. In this way, we are better able to accommodate the solution to the customer's need and even evolve with the customer to new concepts.
We also remain adaptable. If for example, we have a client who wants to scale up the solution already implemented, our internal structure provides the ability to build and enable the solution desired by the client.
The algorithm code implemented by us and in use for biometric services has a high level of trust. However, while the basic function of recognising a face remains the same, the system must be continuously improved to be more and more confident, noticing more differences in the face, more cardinal points, and more processing power.
As technology advances across the board, the tools available to bad actors are also evolving, giving them a greater capacity for systems spoofing, and impersonation at an individual level.
The evolutionary process is therefore critical for algorithms used in the detection of life. These systems ensure that, outside of a controlled environment, the person who is there at that moment performing that operation is an actual living being, not an HD video or a digital morph. That's why it's very important for a liveness algorithm certified today to be certified every year.
We actively participate in measurement algorithm efficiency evaluations, like NIST and MdTF. The tests are conducted in an agnostic manner and at a highly demanding virtual level. We recently have been accredited as top-ranked at the US Department of Homeland Security’s MdTF Biometric Rally 2021. This proof of value validates that the algorithms are more effective in their performance than others. Our engine delivered top performance across all capture devices when excluding the acquisition errors.
But the bulk of our expertise lies behind the veil. The biometric device technology in Vision-Box has machine specialists, mechanical equipment specialists and biometric equipment specialists. We help to enhance the technical capacity of our touchpoints to guarantee an improvement in the capture conditions - whether for fingerprints or face capture. The goal is always that the devices work in the best possible way, with safety, comfort, and confidence.
What makes us bring value to the industry and our customers is the fact that we are critical, that we test, look for alternatives, challenge ourselves internally, and accept external challenges to reach a higher level and train the people who work with us. This is all to provide the best and more secure solution to our clients and users.
Publish date: September 2022