Identity verification through facial biometrics has many advantages, both for the users and companies. Thanks to identity validation solutions, customers can carry out various procedures from any location and without having to go to an office and simply using a mobile phone or any other electronic device fitted with a camera. The software oversees the verification, in a fully automated manner, that they are who they say they are, eliminating unnecessary waiting times.
Benefits for companies are also very evident. By carrying out the validation with biometrics, process security is increased, fraud is reduced, and proof is gathered for posterior claims. Moreover, as it is automated, waiting times and the risk of losing operations are reduced.
But, how do IDDILIGENCE’S facial recognition for identity validation solutions work?
Biometrics are automated techniques that uniquely recognise one person based on one or more proprietary and specific traits of each individual.
In order to carry out this unique automated recognition, biometrics applies technology to generate an algorithm based on some unique physiological or behavioural characteristics. The following are the most currently used characteristics for biometric solutions:
- Digital fingerprint
- Facial biometrics
- Voice biometrics
In this article we are going to focus on describing how facial recognition for identity validation solutions work on a person, like the one IDDILIGENCE by ADDALIA uses.
HOW FACIAL RECOGNITION FOR IDENTITY VALIDATION WORKS
To inexpert eyes, identity validation through facial recognition could look like magic: a simple selfie or video allows the computer solution to verify that the individual is who he/she says he/she is in an automated manner and in real time.
In order to validate identity, biometrics is based on mathematical models, which are the algorithms that are generated according to some input data. In the case of facial biometrics, the algorithm is generated through the analysis of the facial patterns of a person that have been gathered from a photograph on the document that certifies that the individual is who he/she says he/she is, for example, a national ID card, Passport or driving licence.
Then, the solution validates that the biometric pattern extracted from the official document photograph, found on a selfie or video provided by the user, validating a certain percentage of coincidences between the individual on the document and the one who appears in the photograph or video.
VALIDATION OF THE OFFICIAL DOCUMENT THAT CERTIFIES IDENTITY
The first step in order to carry out identity validation through facial recognition consists of the validation of the certifying official document. To do so, solutions like IDDILIGENCE use technologies like optical recognition that analyse the elements that they are made up of to identify, in an automated manner, that it is really an official document.
Each solution supports some types of documents. In the case of IDDILIGENCE by ADDALIA, thanks to machine learning and Artificial Intelligence, the application is capable to automatically identify and classify national ID’s, Foreign ID’s, Passports, EU Citizen Registration Certificates, income statements like tax returns or salaries, bank receipts or standing orders, among others. Moreover, IDDILIGENCE can be configured and trained to automate identification, extraction and validation of specific data and documents, incorporating the additional validations requested by the client company.
YOU ALSO MAY LIKE...
Mobile identity verification: why banks should bet on digital customer onboarding solutions
BIOMETRIC PATTERN EXTRACTION, THE DATA PATTERN AND IDENTITY VALIDATION
Once the solution has identified, classified and validated the document that certifies identity, it analyses the photograph that comes with it, measuring the distance of specific traits, among other factors. On the basis of these parameters, it applies an mathematical algorithm that generates a biometric pattern that is used to validate identity.
Along with the biometric pattern, the software extracts the name and surnames of the document owner, to which a mathematical algorithm is also applied to create an MRZ (Machine Readable Zone), a unique code that is created according to the attributes of the name, surnames and ID number of the person.
Once this information has been extracted, to validate the identity, IDDILIGENCE verifies that the biometric pattern and the data pattern (MRZ) of the certifying official document are the same as the ones found in the other data and documents provided by the user. This means that the software, in an automated manner, validates that the name, surnames and ID number input coincide with those that appear on the identity document, and that the person who appears in the selfie or video is the same person who appears in the photograph.
This result is shown as a reliable percentage that indicates the degree of correlation between the biometric pattern and the individual who appears in the selfie or video. The client company can configure the minimum percentage to accept the operation and consider the identification as valid.
Security levels in identity validation through facial biometrics.
When validating a person’s identity through facial biometrics, there are different security levels that are established in accordance with the type of image requested from the user to verify the biometric pattern extracted from the certifying document.
Biometric pattern validation security levels
Level 1: Selfie. Only one photograph of the individual is required and, therefore, the security level is basic.
Level 2: Video. The user is asked to provide a video file to offer a greater security level when validating the biometric pattern.
Level 3: Life certification. In this case, the user must provide a video carrying out a specifically indicated movement.
Depending on the solution used, one security level or another is available. IDDILIGENCE offers biometric identity validation with the maximum degree of security that currently exists in the fight against fraud (level 3).