In June, Elenium, was asked to be part of a prestigious Biometrics Symposium at the FTE EMEA along with Thomas Hoff Andersson, COO at Bengaluru International Airport, Joe Leader, CEO of APEX & IFSA and moderated by Harald Jentsch, Head of Airports & Baggage Screening at Smiths Detection. We discussed seamless travel and how biometrics will enable the utopian vision for passenger experience at the airport.
Thomas Hoff Andersson spoke about how the Bengaluru International Airport will deliver the largest full deployment of end-to-end biometric solutions at an Asian airport. Joe Leader highlighted APEX’s view on the significance of biometrics to the travel experience by proposing a single token end-to-end biometric ID, which would ease passengers flow through the airport. Prioritising biometric security, Leader highlighted that passenger biometric data must be individually encrypted so that the information would be useless, were it ever to fall into the wrong hands.
Elenium is not a biometric company, our focus is on common use and in particular self-service common use. Our mission is to improve common use self-service and how it is currently implemented in the aviation industry, as compared to other sectors of the economy.
If we are not a biometrics company why were we asked to present at a Biometrics symposium? Easy, our experiences at Hong Kong International Airport and with Amazon Web Services.
One of the difficulties of introducing biometrics into the airport environment is common use. How do you make sure that you are able to retrofit the infrastructure that the airport has recently installed and ensure that the airlines do not have to make massive changes to the application in order to use the technology. This was our challenge at Hong Kong International Airport, where we have 120 kiosks and the location for our Biometrics proof of concept. Our hypothesis was that enrolment of a passengers facial biometrics and generation of a single token could occur at the common use (CUSS) Kiosk by retrofitting a facial recognition camera and registration system without the requirement for an airline to modify their CUSS check-in applications to support the enrolment process.
By intercepting the messages between the CUSS app and Platform and using virtualized devices we could manipulate the workflow for passenger enrollment without changing the airline application
The folks at FTE asked us to open up about some of the lessons learnt while implementing this biometrics solution.
So what happened during our proof of concept? One thing for certain is – we proved our hypothesis. We were able to adapt our platform and onboard two airlines, for group and individual passengers without making changes to their platform. Sounds good, and what did we learn? For one, group bookings are tough when it comes to biometrics. Without going into too much technical detail, the kiosk needs to match the passengers who have had their biometric data taken to their passports as a group. Algorithmically, this is not easy but we were able to achieve a 92% reconciliation rate while enabling the passenger to scan their passport only once during the check in process.
We partnered with Amazon Web Services to create a concept for a seamless travel experience using cloud based biometrics (with no local infrastructure) and a passengers mobile phone for enrolment and as a companion application.
Again, the concept and our hypothesis was proved. We were able to demonstrate a concept where a passenger can check in online and enroll their biometrics by taking a selfie. Using a simple web service by Amazon Web Services we could enable self service touch points with biometrics with a low cost camera. No local infrastructure was needed, we could use the internet through a wireless connection and verification times were less than one second, additionally, we could achieve on the move recognition.
We learnt some interesting lessons in this process too. Mobile applications need to use computer vision to sanitize the background, cutting out “noise” and focusing on the “selfie”. Artificial intelligence determines the correct point of focus. The biggest challenge was ensuring the “right” person was recognized and not people walking past or in the background, we used proximity sensors to improve this. Interestingly we detected a racial bias in biometrics. This would be reduced over time and can be minimized by improved training sets. We will soon be implementing this solution, we call Voyager, in the Middle East.
It was a real pleasure to share the stage with Thomas, Joe and Harald and share our thoughts on how we can enable truly seamless travel by leveraging biometrics. The Summit generated strong interest and it was good to field many questions on Elenium’s experiences. We look forward to rolling this out to other airlines and kiosks beyond Hong Kong International in the near future.