Biometrics is the study of people's unique physical and behavioral features via measurement and statistical analysis. Biometrics includes not just physical qualities such as fingerprints or facial features, but also behavioral characteristics such as walking form or breathing patterns. Biometric technology uses these biometrics to identify a certain person.

The most pressing concern of biometrics is whether they are substantial enough to distinguish people based on them. Some biometrics, such as fingerprints, have previously been established to be unique to each individual, while others, such as breathing patterns or walking styles, are still up for debate. Many individuals were initially hesitant to use facial traits to identify persons since facial features may readily alter owing to diet, spectacles, hairdo, or even cosmetics.

 This is why biometric technologies make use of AI and deep learning. Traditional cryptography merely verifies that the user has input the correct password: the two passwords must be identical or distinct. When it comes to biometric authentication, however, there are no hard and fast rules. As a consequence, rather than just remembering the correct password and comparing it, biometric technology collects hundreds of data points and searches for patterns or similarities. In the case of facial recognition, the computer takes numerous photos of a person's face and analyzes them to see which features are more beneficial in recognizing the person than others. Skin tone, for example, may be less relevant information because it fluctuates depending on the amount of light available. The computer then constructs an activation function, which is a formula that informs the authentication outcome.

The beauty of AI and facial recognition is that the formula is created by computers rather than people. Computers, like live beings, consider each attribute and make their conclusions.

김강민 강남포스트 학생기자  webmaster@ignnews.kr

<저작권자 © 강남포스트, 무단 전재 및 재배포 금지>