/Facial Recognition and Facial Authentication: Let’s Cut the Confusion

Facial Recognition and Facial Authentication: Let’s Cut the Confusion

Facial Recognition

We are all aware of the concept and usage of biometrics for Facial Recognition and Facial Authentication. With the ever-evolving deep learning algorithms, detecting face and comparing with the stored face records is achieved in fraction of seconds. Face recognition and face authentication seems similar but has varied meanings, applications and need to be classified with growing advancement in use cases.

What is Facial Recognition?

Facial Recognition technology has now been around for years and, as we discussed the ever-evolving technological advancements, it is not limited to just humans. Its application has expanded and it can identify any individual on earth (even cats and dogs).

One of the biggest uses of Facial Recognition technology is in law enforcement world over. Many state police departments have been procuring such sophisticated face recognition systems and integrating with their CCTV network in their city surveillance projects to capture crime proactively. Indian Railways has also started installing such systems in their stations to identify the criminals and missing persons. Earlier this year, one of the biggest temple complexes in India installed Face detection systems with associated analytics to manage the devotees’ progress.

Faces detected in this process mostly from CCTV network feeds are matched with large number of face database to find the match. Such searches in technical terms are called ‘one to many’ verification. NIST (National Institute of Standard and Technology), the largest testing agency in the US famous for testing AI based products and solutions, uses stringent parameters to test face recognition solution made of 1: N environments. Apparently, for Face Recognition systems one needs large processing power like GPUs to get better results of recognition accuracy.

What is Facial Authentication?

Most of us today are connected to the world via mobile phones. We have literally everything on our phones – from saving passwords on our phones to online banking transactions. We shop online, we pay online, and have everything that is personal to us saved on phone-based online mediums!

Since everything personal to us is available on hand-held device, it is easier for a hacker to hijack your identity and rob you off of your wealth and/or threaten to expose your personal details that might put you and your family and/or loved ones in danger. This is where the Facial Authentication feature plays a powerful role.

When an online user creates an account, his/her profile is vetted through some identity proofing or verification process. For this, they are usually given a unique credential (i.e. username and password combination) that lets the user access their online account and perform certain actions online via his/her account. This verification of credentials is what we call authentication.

Now, Facial Authentication is a component of biometrics that works as a “verification of credentials” of an account-holder via Facial Recognition. Here, face matched with its own (1:1) database is so easier to authenticate. While face based authentication is safe, it is not foolproof as the same works on two-dimensional image and a photo or video of person can be replaced with actual image to get the authentication cleared. To make this process secure, a new feature has been introduced, called liveliness detection. Liveliness detection is used in face authentication process to ensure that the image presented is of a live person and no photo or video is being used to hack the system. One of the most popular ways of liveliness detection is ‘pose based liveliness detection’, wherein a user is asked to turn left or right or asked smile for few seconds to ensure the person presented is alive. However, further advancements are being explored in AI algorithm where a photo or video of a person can be easily differentiated with his live image.

We’ll see more robust face-based authentication system in coming years which will be foolproof from the frauds and online hacking.

So, a one-line basic difference between Facial Recognition and Facial Recognition is:

Facial Recognition is a software application that uses your facial features and contours to analyse, compare with stored face database and identify an individual, whereas Facial Authentication is a form of biometric authentication that uses an individual’s stored biological biometric template and other details to match him from his own details. However, liveliness detection is a key to the success of face authentication use cases.

Bikas Jha
The author is Country Head, RealNetworks Pvt. Ltd

Leave a reply

Your email address will not be published. Required fields are marked *