Overview of feature

 

To answer the question: “Is this a live person?” IDUX Flow uses passive liveness detection.

Passive liveness is considered a low-friction approach for user experience as the technology works with a single image into the authentication process to exam and detect attacks. 


How it works

 

Liveness detection works by examining a single frame through a biometric system to measure and analyze physical characteristics in order to determine if a biometric sample is being captured from a living subject who is present at the point of capture. 

What enables the use of a single image to perceive liveness is an artificial intelligence-based approach. The single image is processed through a sequence of neural networks. Each one diagnoses a particular element of the frame to spot artifacts that help differentiate between an image of a live person and a presentation attack. The system then considers the results of these neural networks and real-world data and provides a threshold to mitigate false acceptance.

The technology detects presentation attacks. These include: 

  • Photo attack: when fraudsters gain access to an individual photo through social media or other and create a 2D mask to mimic the individual’s image.

  • Deepfake: when fraudsters through editing create a realist version of an individual. 

  • Model or 3-D Mask: when fraudsters invest in 3D models to mimic an individual.

 

Advantages of passive liveness

  • Low-effort: requires no action by the user.

  • Occurs when the user takes a selfie.

  • Based on single image.

  • Fast approach due to its low-friction.

  • Passive approach is ISO 30107-3 compliant.



Step-by-step

  • User takes a portrait photo/selfie for liveness check.

  • Artificial intelligence analyses the image for liveness.

  • The system returns a liveness probability score.

 

Photo requirements

A few requirements must be met while providing a portrait photo for verification:

  • Good lighting and absence of effects and filtering on photos

  • Avoid sharp glare, like from overhead lighting and direct sunlight

  • Avoid blurring image due to motion during capture 

  • Avoid wearing glasses and hats

 


Where can you access liveness results on IDUX platform

 

Once logged in the platform:

Go to: Verifications > Select an individual verification > Overview > Biometrics > Face Liveness

As a result of deep neural network analyses run on the image, a threshold is translated into a successful or failed result.


Note: Biometrics result such as Face matching should be evaluated in conjunction to other modules results in favor of meeting your company’s individual criteria for identity verification.


 

While data is visible on the platform, to view and analyze the image provided for liveness:

Go to: Verifications > Select an individual verification > Uploads > Portrait photo

 

 

This data is also available in PDF format and API. To export the verification in PDF format:

Go to: Verifications > Select an individual verification > Export PDF