Ever wondered what is the difference between Face Recognition and Face Detection? Well, you’re not alone. In fact, in our industry, we find that face recognition and face detection are two terms that are commonly misunderstood and as a result, tend to be used interchangeably. If you do a quick Google search, you’ll see exactly what we mean. In this article, we will shed light on these two technologies and give you an overview of what they entail.
What came first, detection or recognition?
In recent years, face recognition has gained a lot of attention and is now appreciated as the most promising application in image and video analysis. Face detection is a significant part of the facial recognition process. In fact, it is the first step towards facial recognition, as well as other processes such as face analysis.
A face detection system uses machine learning or deep learning algorithms to determine the presence and location of faces in a still image or video frame. The process is anonymous – face detection is not used to identify the detected face or faces. It also does not store any personally identifiable information. It merely places a bounding box around the faces that it detects to show where in the image or video those faces are situated.
What is Face Recognition?
Face recognition, as you might have gathered by now, is a biometric technology that does more than just detect a human face in an image or video. It goes a step further to establish whose face it is. A face recognition system works by taking an image of a face and making a prediction about whether the face matches another face or other faces in a provided database. The technology is designed to compare and predict potential matches of faces regardless of their expression, facial hair, and age.
With recent advancements in deep learning, face recognition has become highly accurate, with accuracy levels of over 99% with confidence thresholds of false acceptance rate of one in one thousand.
How businesses use Facial Recognition to enhance their processes
Face recognition technology is a valuable tool for those who need to verify identities and authenticate users. Financial institutions have adopted facial recognition to streamline KYC processes and provide a frictionless banking experience for their customers. Other applications for this technology include payment verification and access control.
Let’s have a look at these applications in more detail:
Know Your Customer (KYC)
Financial institutions such as banks, insurance companies, asset managers, and payment service providers must comply with strict anti-fraud and anti-money laundering (AML) laws and regulations. And part of this compliance involves knowing who their customers are, as well as determining whether the people performing financial transactions really are who they say they are.
With the rise in digital and online channels in the finance sector, being able to perform KYC remotely is becoming more and more important. And this is where facial recognition comes in. By using facial recognition software in their KYC processes, financial institutions no longer need to ask their clients to be physically present to open an account or to perform high-value transactions.
Instead, they can ask clients to send a copy of their identity document along with a selfie to authenticate their identities during onboarding. The facial recognition software then compares the selfie with the photo in the ID to determine whether there is a match. It is also possible for the software to determine whether the selfie was taken of a real person at the time of submission, or whether it is likely to be a spoof (liveness detection).
By now, most people are used to two-factor authentication when making electronic payments – whether it’s having to provide a CVV number from a credit card or having to provide a one-time pin sent via SMS when making a purchase online, or having to enter a pin code when swiping a bank card at a point of sale. Thanks to technology such as Apple’s Touch ID and Face ID, people are now also used to authenticating online transactions using their biometric data.
This is fortunate, since the PSD2 regulation that’s coming into effect in the European Economic Area (EEA) in September 2019 requires two-factor authentication, with biometrics being stated as one of the authentication options. This means that an increasing number of payment service providers are going to be considering technology such as facial recognition for verifying payments.
There are multiple ways that payment service providers could implement this in practice. They could, for example, give customers the option to use the Face ID feature on their smartphones, or, in an in-store environment, they could install cameras that are able to identify faces at the payment terminals.
Whether it’s used for accessing smart buildings, securing devices and apps, streamlining check-ins at airports and hotels, or admitting attendees at events, facial recognition can help to improve and speed up all these processes. Depending on the use case, it can contribute to better user experience, quicker processing time, and increased security.
Similar to payment verification, face recognition for access control can be implemented using the built-in smartphone functionality for Face ID, or by using facial recognition cameras.
Face detection is the first step and a part of bigger computer vision processes such as face analysis and face recognition. Face recognition is a more complex process that starts with face detection and continues to establish whether or not two or more faces match, usually for the purposes of authentication or identification.