MATICLASS: Understanding Liveness Detection

Share on facebook
Share on twitter
Share on linkedin

Hello, my name is Val and this is my MatiClass. Today I will teach you all there is to know about Liveness Detection; also known as Proof of Life in LatAm. Now without any further ado let’s get to our lesson.

In biometrics, Liveness Detection, also known as Proof of Life, is a test that corroborates a live person is present. In other words, making sure a live human being is verifying himself and not a mannequin. 

Liveness is often used along with Facematch recognition since usually, neither of them alone is sufficient on its own. Facematch recognition will happily accept a photo of a user since it only checks that the faces are similar. Liveness detection will accept any live person since it doesn’t check face similarity. Therefore it’s advised to use both.

What are the different types of liveness checks you can perform?

Liveness is carried out through different algorithms that analyze data collected from the video the user uploads. This in order to determine whether the source is physically present or a spoof (spoof is a term used in our industry to say imitate, hoax or trick).

There are two main categories of liveness detection:

Active: Active facial liveness is the most common option used in today’s biometric systems. With it, users are given tasks such as: “making two circles with their head,” and the system waits as they complete the challenge.

Passive: Passive liveness detection, on the other hand, uses algorithms to detect indicators of a non-live image without user interaction. The capture of high-quality biometric data during enrollment improves the performance of matching and liveness detection algorithms.

One or the other may be preferable in specific scenarios, but they generally work better together.

So now let’s take a look at how the Active Liveness test validation process actually works:

As stated above the active validation process consists of a series of analyses that are executed at the moment the user uploads their liveness test video. Users are asked to record a video, approximately 6-7 seconds in length, in which they can be seen making two circles with their nose.

In short, our validation process is:

1) The video uploaded by the users is converted into a set of frames, then our algorithm determines which ones contain a clear human face.

i.e. :

2) Once the user’s face has been located, a virtual, 3D structure of the face on each frame is analyzed.

i.e. :

3) Our system determines if the observed movement corresponds to the natural movement of a human being. This means it complies with the following:

  • It presents no discontinuities.
  • There are no sudden movements.
  • The movement is between the range of normal muscular motion.

If it’s determined that the subject in the video does not meet the threshold defined on your Mati dashboard standard; it is declared a failed liveness test. Otherwise, it is determined there’s a real human being in the video.

Why and how should you use LIVENESS DETECTION at account creation? 

Liveness detection is useful not only for authentication but also to verify your identity. Besides confirming that the user is the same person who initially enrolled, Liveness can also be performed as part of an onboarding process to confirm that the applicant is, in fact, a real person. An example would be: using a neobank app to apply for a new account. The person is not known to the bank, or present at the institution to corroborate, so liveness detection can be used to confirm that the applicant is not trying to open a fraudulent account. Today’s technological advances can be exploited by fraudsters easily therefore, liveness detection is a vital feature for modern biometric-based recognition solutions. Reach out to one of our Liveness Detection experts to learn more.

Share this post with your friends

Share on facebook
Share on twitter
Share on linkedin

Leave a reply:

Your email address will not be published.

Subscribe to our Newsletter