Authentication, identification and their characteristics

To better understand the possibilities of facial biometrics, today we will focus on the differences and similarities between digital authentication and identification. Because, although we may think they are interchangeable, they are in fact different ways of keeping the process safe and efficient. 


In this case, we are responding to the question: is this person who they say they are? To answer this question, a comparison will be made between this person, who must be in front of a camera, and a previously stored photo of the same person.  

It is also known as 1:1 because the facial pattern we evaluate is only compared to the template of the same user. This requires a good facial pattern extractor, which will later be used by the matcher, the technology in charge of checking whether the patterns belong to the same user. If correct, the matcher will give a positive result. This procedure can be used, for example, by banks that need to verify the identity of their customers in order to give them access to their accounts. However, it can also be applied to any other sector that requires digital authentication. 


This technique answers the question: which person is it? It answers by comparing a person’s biometric data with that of a set of users. The aim is to find the identity of this user within the group.  

We also refer to it as 1:N because the identity must be located among N number of patterns with different identities against which it is to be compared. In this case, the patterns of the various profiles should be as different as possible. A company can use this method to record the clocking in and clocking out of its employees. Or to let a season ticket holder into a stadium to watch a match. 

Both, authentication and identification, work with enrollment, because user data is needed in order to be able to make the comparison correctly.  Attributes such as caps, glasses, scarves or masks should be taken into account, otherwise the pattern of this verification system will not be accurate and could cause problems. In this case it is also essential to have an inclusive technology that controls the level of demographic bias (link to last post), which can also lead to a negative result.