[Explanation] Recently, a digital blogger broke the news that the area collected during face recognition is not only the head displayed on the screen, but the entire range covered by the camera.

  In response to the question about the collection area of ​​face recognition, a researcher at the State Key Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences, gave the answer during the interview.

  [Concurrent] Researcher of the State Key Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences impressively

  Face recognition itself only needs to review the range of faces, but in the process of face recognition applications, for example, when using a web camera for face recognition registration, a larger collection area is often required.

This is because in face recognition, it is necessary to confirm that there is a real person in front of the webcam, rather than a photo or video of a person's face. This is also called live detection or face verification in face recognition.

  [Explanation] It is surprisingly introduced that the counterfeit verification in face recognition requires a larger collection area, and at the same time, there are generally reviewers in the background to judge the face recognition process.

  [Concurrent] Researcher of the State Key Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences impressively

  Face authentication mainly judges whether the face in front of the camera is artificially forged. From a technical point of view, a larger collection area is required at this time.

Therefore, when face recognition, especially the first face registration, there is generally a review process to determine whether the person in front of the camera is the real person of the applicant, for example, we need to determine whether he is taking someone else’s photo. For example, a celebrity photo to register.

  [Explanation] Nowadays, face recognition technology is more and more commonly used in all walks of life.

It also reminds the public to pay more attention to the abuse of face recognition technology in life.

  [Concurrent] Researcher of the State Key Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences impressively

  In some public safety areas, the use of face recognition technology is a normal situation.

But for example, during the use of some mobile apps, forcing users to sign data usage privacy agreements, or even collecting facial image data without informing the users, this is an abuse.

There are also many community access control or shopping mall monitoring, using face recognition without informing the owners or consumers, this is also an abuse.

In addition, in other public places with surveillance cameras, if the face recognition area is not marked, or face data is automatically collected, this is also an abuse.

  [Explanation] While facial recognition technology facilitates people's lives, the safety and stability of its technology remains to be seen.

Does face recognition have to "put on clothes"?

How to better standardize the development of face recognition technology?

Awesomely give advice on this.

  [Concurrent] Researcher of the State Key Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences impressively

  On the one hand, the use of facial recognition should be strengthened, and Internet manufacturers and facial recognition technology providers should be regulated to avoid illegal use of facial image data; at the same time, user privacy agreements should be regulated to avoid manufacturers from forcing users to provide image data use rights .

On the other hand, for scenes that require face recognition, you can mark and prompt the user to perform face recognition and image collection to ensure the user's right to know.

Finally, it is necessary to strengthen the popularization of science and raise the public's awareness of privacy protection.

  Reporter Xu Pengpeng reports from Beijing

Editor in charge: [Ji Xiang]