Imagine the following situation: you want to open an account in a digital bank and opt for the convenience of doing this through application. In one of the onboarding steps, the app requests an image of your face and describes instructions for an acceptable capture. The capture ends up not coming out according to the instructions, however, the image is accepted.
With the account created, you'll, excited, organize your finances, and when you try to access it using face recognition, it doesn't work. You try other times and the error remains.
What was supposed to be a practical and easy solution becomes a problem and the insistence of the error brings the feeling that your data may have been defrauded since it is not accepted.
Learn about the top security flaws tendencies for 2022 here.
If the situation presented above already causes discomfort, when brought to an extreme case, such as a crime scene, the impact is even greater.
In this example, a suspect commits a crime and is caught briefly by a monitoring camera. From this fragment, it would be easy to identify the subject by crossing the detected fragment with the forensic database, right?
Wrong. Identification would only be possible if the registered image was of sufficient quality to "match" the existing fragment. In short, several identification errors happen when biometric cadastral qualification is not made successfully.
Whether it's the difficulty of accessing the bank account or the possibility of a crime suspect being released, the lack of biometric quality can bring irreparable damage.
Thinking about it, Ph.D João Janduy, BioPass ID Researcher, in his doctoral thesis, developed a proprietary algorithm with the intent to optimize and analyze the quality of biometric images.
Before we meet him, let's first understand what biometric qualification is and what is its importance. Follow the reading!
Understanding Biometric Qualification
The biometric recognition process happens in 3 main steps: capture, extraction and identification.
Capture is its first step. Since it is the basis of the process, its effect directly impacts the consequent processes.
In turn, extraction is the step in which the minutiae of biometric image attributes, that is, the characteristics that make biometrics unique, are collected.
Finally, there is the identification stage, in which biometrics will be compared with each other and thus identified.
It can be seen that one step is directly linked to the other, making it clear that when one of them fails, the others will respond according to the error, so if the capture fails, the identification tends to fail as well. This causes frustration for those who use it and losses to those who provide it, given that each print has a cost.
These failures originate at the moment of registration, the so-called onboarding. It is at this moment of onboarding with biometric capture that qualification comes into play, to increase security in a system and reduce the error rate, making identification clearer and more effective.
When done in the face-to-face mode, onboarding has a person responsible for inducing the quality of registered biometrics.
In digital onboarding, it is an algorithm that is responsible for inducing the success through instructions to acquire a quality biometrics. In this case, the instructions by themselves do not solve the errors of a capture.
For this, there must be a system with an algorithm capable of measuring and pointing out the registration errors, significantly increasing the chances of success.
Quality API
Understanding biometric qualification as a biometric recognition facilitator, Quality API has become another BioPass ID solution, bringing speed, security and preventing future errors that bring so much dissatisfaction.
It is an API package with features that accuse and investigate biometric capture, and allow the qualification of registration data, either by facial recognition or fingerprint.
In addition, it has security levels that measure the required quality according to the customer's discretion.
To achieve standardization in fingerprint qualification, the algorithm acts according to NFIQ (NIST Fingerprint Image Quality), a software that allows the worldwide standardization of fingerprint sensors with image qualities, belonging to the FBI.
For face recognition, the Quality API is based on the international standard used for issuing passports, the ICAO.
It uses 27 attributes to check biometric quality with some requirements such as not looking away, not being a blurry image and not using accessories, requirements that you probably have already seen in some registration.
Overall, these measures are required to make verification safer with the increased quality of biometric prints.
Conclusion
It was noticed that biometric qualification is not only present in various fields of our daily lives, but it also has a critical importance, since it carries with it identity marks from all over the world, besides serving as a solution to foster a passwordless future and even be able to promote justice in criminal cases.
Find out here if the future is passwordless.
Investing in biometrics without worrying about quality is like doing a job in half. The completion of the biometric process is done with the participation of the qualification.
Test our solutions and ensure cutting-edge biometric technology in your business!
Access our demos!
