Smart and secure travel: how and why apply facial biometrics in airports
Before the pandemic destroyed our way of life, global passenger traffic used to rise and force air transport hubs to invent new methods aimed to increase airport capacity. One of such methods is to replace passport and boarding pass by biometrics at all the journey stages from check-in to boarding. This article covers using biometrics for passenger verification & identification within transport hubs and explains its benefits.
How do biometrics identification & verification work?
Biometrics identification is the process of recognizing a person through the facial image. It is advantageous to both users and businesses as using biometrics cuts hardware expenses, reduces the time of customer service, prevents fraud, automates business processes and increases profit.
The main purpose of biometrics verification is to make sure that the human in front of a camera is the very that person he pretends to be. So the first step of this process is to compare a passengers’ passport with his image which can be done at the reception during self-check-in. Meanwhile, Svort neural network learns to recognize this particular passenger and transfers his biometrics data to all other touchpoints. Since this moment all the passenger needs for passing the journey steps is his face.
It is possible due to a particular neural network algorithm combined with infallible liveness detection that determines whether it interacts with a human being and prevents spoofing attacks employing photo, video, or mask. Although our liveness detection tests excellent we reiterate that its efficiency also depends on hardware characteristics like computer or webcam specifics, depth sensors in the camera, etc.
What are the benefits of Svort biometrics and what tasks does it resolve?
Unlike existing systems with one or several standards, Svort basically has no standards/references (photos or other human-interpretable data on face parameters). Svort is the 3D reconstruction of the user’s face, which is irreversible, non-human interpretable, and anonymous. It also recognizes passengers in the most precise manner even from those angles where current systems are experiencing difficulties.
No sensitive data like facial images are stored or transferred. Svort stores and transfers only neural networks that are trained to recognize passengers. These networks are depersonalized and anonymous as they do not contain any data on passengers and it is impossible to clarify some person’s face features (like eyes color, face geometry, etc.) using it. Such data can be transferred to any transport hub as it does not contradict GDPR.
Svort is not just a biometric identification system, but also the liveness detection that mitigates face-spoofing and is able to take advantage of cameras capable of constructing depth maps. Yet, Svort can with the same effect perform all its functions using GB/smartphone/notebook/etc. cameras. Svort does not have strong requirements for hardware and can be scaled horizontally if necessary to increase productivity. It is backward compatible with installed systems of both video surveillance and access control.
Svort is faster, more reliable, and safer than other biometrics solutions with facial recognition. Neither sunglasses, beards, and makeup nor face masks would not significantly impede its functions.
How Svort neural-biometric verification and identification work
Biometrics verification can be integrated at all steps where using a passport or boarding pass is supposed: check-in desks, bag drop units, gates, boarding, etc. The main idea is to replace documents with passenger’s face once his identity was verified. Facial biometrics can also enable airports to create lists of suspicious people or criminals and swap these lists to all transport hubs. Let’s clarify in detail how exactly neural-biometrics identification occurs.
Online neural-biometric registration for a flight
Actually, such registration can be used for everything that implies buying and checking tickets. After completing registration (on the website or in the application) a client is offered to pass neural-biometrics registration.
- A user needs to provide his ID and complete the registration process on his smartphone. For this purpose, he will be offered to pass a rotational challenge. He will also have to escape a virtual maze to confirm liveness.
- During the rotational challenge, the camera scans the passenger’s face comparing the image with the photo in ID, building its 3D face model, and training neural networks to recognize this passenger.
This action can be performed as a one-time registration or issuing a kind of biometrics ID. During these processes, Svort extracts some non-human interpretable vectors. Using a neural network, Svort converts these vectors into a biometric key realizing the concept “You Are The Key”.
Actually the key is a neural network that is trained to recognize a particular user. It is irreversible that makes it similar to password hashes. This allows to store and transfer it according to the same requirements as exist for hashes. Note that the neural network is neither a reference image nor other data that would provide specific characteristics of a registered user’s face.
After the neural network is saved in storage, it can be used for subsequent identification, authentication, and verification of a previously registered user.
Offline neural-biometric registration for a flight
For offline registration, a passenger uses a terminal or camera at registration desks in the airport instead of their own device. There are two ways of biometrics registration at the airport.
- The video stream is continuously transmitted to a data center where it is processed. It creates a significant load on the network.
- In the case of local processing, already trained neural networks or biometric vectors are transferred from the terminal or the camera at the registration desk to the data center. It significantly (up to 99%) reduces traffic and load on the network.
One more peculiarity of offline registration is that using public terminals instead of personal devices leads to either queues or low quality of biometrics registration. The way out is the dynamic registration, that completes neural network training. Dynamic registration collects missed biometrics data while a passenger moves through the airport until the neural network won’t be trained to recognize this passenger (or until another trigger emerges).
In summary, Svort biometric solution can:
- improve analytics of passenger traffic and customer service;
- help monitor overburdened areas within transport hub and reorganize its work;
- enable airports to make personalized offers to passengers in duty-free stores and service desks;
- increase the safety of air travel with the support of international crime control organizations (like INTERPOL, Europol, etc.) by comparing passengers with the databases of these organizations, identifying unwanted persons, and sending notifications to the staff.
Although air carriers are currently experiencing challenging times due to the pandemic, after everything is over and the borders open, airports will experience even greater load. That is, they will have to adapt, but increasing bandwidth will not be enough. The touch-free trend is gaining momentum and public places are already starting to implement the technology to protect people from diseases spreading through human-to-surface contacts. Here you can learn more about how Svort implements touch-free technology.
Svort is faster, more reliable, and safer than other biometrics solutions. We are open to discussion, collaboration, and new opportunities.