What will you see in the background during facial recognition?

Unlock the screen, log in to the account, pay, check in.

Unlock the screen, log in to the account, pay, check in... The moment we wake up and pick up our phone, we start the "face brushing" activity. From morning to night, we may perform facial recognition in any scene, but when alone, we need to pay attention to "brushing our faces".

Credit: huimeng@TOPYS

More than just facial recognition

Everyone is familiar with the face recognition interface, and all other parts are covered except for the area where the face is displayed.

In fact, the range collected during facial recognition is not only the circular area, but also the entire shooting range of the front facing camera. Given that some facial recognition information may be manually verified and retained information may be uploaded, there is also a Weibo hot search reminding everyone not to perform facial recognition without wearing clothes.

Comparison between front and rear end of face recognition

A friend who just learned about this may sweat a cold sweat, but some experts say that in order to save bandwidth, except for special scenarios such as banks that require uploading videos for manual verification, most smart video cameras do not upload videos, only return characteristic data, which cannot be restored after desensitization.

The requirements for handling sensitive personal information in the "Information Security Technology Personal Information Security Specification" (GB/T35273-2020)

How does facial recognition recognize us?

In face recognition, the machine first needs to locate the face, extract the feature items of the face, and then compare them with all known faces to determine the identity of the face. So how do machines achieve these things?

We have portrait mode when taking photos with our phones or cameras, which can easily detect the position of faces and help the camera focus quickly.

Specifically, the machine will analyze each pixel in the image and its surrounding pixels, and draw an arrow based on the brightness. The arrow's direction represents the direction in which the pixel gradually darkens. These arrows are called gradients and can display the process of an image flowing from bright to dark.

With these arrows, the algorithm divides the pixels into larger regions and extracts local features of the image based on the frequency of arrows appearing in that region, thereby recognizing the face.

Finding a face is just the first step, and the machine still needs to recognize which face it is. Firstly, the algorithm will roughly align the faces of 68 people based on the common points on their faces. These points are like coordinates, and even if the angle is different each time, the machine can recognize them.

Afterwards, the algorithm will further analyze the image and obtain 128 feature measurements. By analyzing different photos of the same person, algorithms can almost always obtain almost identical values, allowing for the search and input of facial sequences with similar faces in the face database.

Facial recognition needs to be standardized

In recent years, the use of facial recognition technology has brought convenience to our lives, but it has also raised issues with personal information protection compliance.

This year's "3.15" evening party exposed the misuse of facial recognition by multiple enterprises. For example, multiple stores in Shanghai Kohler Bathroom have installed cameras with facial recognition function, and customers will be secretly captured facial data without knowing it. The stores will leave records of which store and how many times customers have visited.

There is no separate legislation in China regarding the use of facial recognition technology, but relevant laws, regulations, and national standards have been gradually introduced, and specialized rules and standards are also being further developed.

GB/T35273-2020

Information Security Technology Personal Information Security Specification

This standard provides detailed rules for the collection, transmission, and storage of personal biometric information, such as obtaining explicit consent before collecting personal biometric information and ensuring that it is a fully informed, specific, and clear expression of willingness.

GB/T38671-2020

Information security technology - Technical requirements for remote facial recognition systems

This standard specifies the functions, performance, security requirements, and security guarantee requirements of information systems that use facial recognition technology for remote identity authentication on the server side.

GA/T1470-2018

Classification of security and face recognition applications

This standard provides classification basis and methods for the development, application, and management of facial recognition application systems.

If relevant explanations are provided on whether there are facial anti-counterfeiting requirements for identity verification, there may be no facial anti-counterfeiting requirements for on-site confirmation of identity verification (such as handling business at a bank) according to security level requirements. For remote confirmation of identity verification (such as opening an account remotely through the network), due to the lack of supervision at the image collection site, there should be facial anti-counterfeiting requirements to ensure that the system is not susceptible to artificial facial attacks.

In addition, national standards such as "Information Security Technology Facial Recognition Data Security Requirements" and local standards such as "Application Specification for Classification of Facial Recognition in Public Places" are being developed, which will further safeguard personal privacy and data security.

Content source: Comprehensive Media Overview

Image sourced from the internet

Disclaimer: The content of this article is sourced from the internet. The copyright of the text, images, and other materials belongs to the original author. The platform reprints the materials for the purpose of conveying more information. The content of the article is for reference and learning only, and should not be used for commercial purposes. If it infringes on your legitimate rights and interests, please contact us promptly and we will handle it as soon as possible! We respect copyright and are committed to protecting it. Thank you for sharing.(Email:[email protected])