Portrait of woman with long brown hair from the side
 
#Identity Technology

Morphing: from impostors to spoofing

New Technology
6 Mins.

Fraudsters are increasingly using easily accessible digital manipulation techniques in morphing attacks. Harder to detect than classic presentation attacks, morphing attacks challenge governments to take tighter control of the issuing process. Meanwhile, increased effort is being put into researching morphing detection using advanced computational techniques.

Remember the hours dutifully spent queuing for a border guard to manually check your passport? Those days are thankfully over in many airports. The use of automated face recognition speeds up airport security checks, making your passage through passport control safer, smoother, and hassle free.

Whether at border control or unlocking smartphones, biometric recognition systems are becoming more widespread and sophisticated. As with any new technology, though, there are challenges: biometric authentication can fall prey to morphing or presentation attacks.

What is morphing?

According to NIST, “Face morphing is an image manipulation technique where two or more subjects’ faces are morphed or blended together to form a single face in a photograph.”1

With morphing, criminals try to trick face recognition systems, like the ones used at border control. During the passport application process they present a photo with a merged image from two or more different individuals.

In this way, two or more people can use a single passport with this photo. In comparison, a presentation attack requires the illegitimate traveler to wear a mask or other means to deceive the automated face recognition system.

Same, same but different

Face of a woman morphed in three steps
Morphing attacks are far harder to detect than classic presentation attacks

Morphing is clearly a technique that needs to be taken seriously. The attack is far harder to detect than classic presentation attacks.

The photograph has to be processed by special algorithms to detect irregularities left by the morphing process. New morphing techniques are in a constant race with new detection methods: continuous research is necessary. Today, artificial intelligence (AI) plays a major role in detecting these attacks.

Of course, the best way to prevent morphing is to make sure that the photo can’t be manipulated when it’s created. However, this is only possible if you’re in control of the issuing process, unlike with foreign passports, explains Frank Schmalz, VP Head of Innovations at Veridos GmbH.

“New morphing techniques are in a constant race with new detection methods“
Frank Schmalz
VP Head of Innovations at Veridos GmbH

EU Research: PROTECT and D4Fly

As one of the world’s leading identity solution providers, Veridos ensures that its biometrics technology is continuously advanced to stay ahead of the game. In this way, it can strengthen European smart borders to aid law enforcement, adds Rafaela Pernice, Deputy Head of PR & Marketing Communications at Veridos GmbH.

To this effect, Veridos is leading an EU-funded research project called D4Fly, working with 18 partners to make border crossing faster and more secure and significantly decrease fraud. State-of-the-art border control solutions under review include biometrics-on-the-fly, document forgery detection, and the benefits of blockchain technology in identity verification.

D4Fly is also putting significant effort into developing new methods of morphing detection with AI. Testing and demonstration will focus on four different border control points and one document fraud expertise center in the UK, Greece, Lithuania, and the Netherlands.

Moreover, Veridos has co-coordinated PROTECT: another EU-funded project on biometrics and no-gate crossing technology. This examines emerging and contactless biometrics and less obtrusive methods of biometric data capture and verification. Other areas under research included “biometric corridors,” through which travelers can pass without having to stop.

  1. NISTIR 8292 Face Recognition Vendor Test (FRVT) Part 4: MORPH – Performance of Automated Face Morph Detection, Mei Ngan, Patrick Grother, Kayee Hanaoka, Jason Kuo, 2020

Published: 18/05/2020

Share this article

Subscribe to our newsletter

Don’t miss out on the latest articles in G+D SPOTLIGHT: by subscribing to our newsletter, you’ll be kept up to date on latest trends, ideas, and technical innovations – straight to your inbox every month.

Please supply your details: