When you select Face Recognition users can make a login using their face, this is one of the most useful tools to avoid spoofing because the user must make their validation in real time.

How this service works?

When the user performs a login process, The Face recognition service does the following:

• Makes a search in the stored users to find similar matches, this to allow the user to gain access depending on if they are already in the DB or not.
• Makes a Liveness detection, that means that Verifik makes a process in which we validate that the image came from a living person, and it wasn't altered in any way.

As you can appreciate in the above image there are multiple options that you need to select, next you will find specification on each of these fields, what they do and a few recommendations of each of them.

Liveness Minimum score

When someone performs a Face Recognition, Verifik performs a Liveness Detection, in this process we check the image provided, then return a score between 0% - 100%, the higher the score the higher are the probabilities that the image came from a living person, and it wasn't altered in any way.

The Liveness Minimum score allow client to specify the minimum % that they are willing to accept in their projects, in Verifik we recommend a percentage of 75% to achieve optimal results, higher percentage require better cameras and better environmental conditions that are not easy to achieve in a normal end user scenario.

Search Minimum Score

When someone performs a Face Recognition, Verifik performs a Search, in this process we check the image provided and then search in our DB for matches, in a login process the only way to make a successful login it's making this comparison.

Verifik makes a comparison in all the current users that you have, an in each of them makes a coincidental percentage between 0% - 100%, the higher the score it's more likely that the person is the same that the one stored in the DB, using Search Minimum Score you establish the minimum score of comparison that you want in order to allow users to make a successful login.

In Verifik we recommend a percentage of 85% to achieve optimal results, higher percentage require better cameras and better environmental conditions that are not easy to achieve in a normal end user scenario.

Define type of search

When we make a search in Verifik’s DB, we have two options in order to make this process, Fast Searching and Accurate Search, ****the main difference in these options is that using Accurate the Search takes more time, but guarantee a better mapping in the existing faces stored.