Recently, with the occurrence of incidents such as “94-year-old man being picked up for facial recognition”, “wearing a helmet to see a house in order to hide from facial recognition”, people’s facial recognition is frequently searched, and the problem of technology abuse has caused widespread concern.
Following this, some rumors also “quietly dispatched”: Any organization of the facial recognition system can install it if it wants to? Can’t perform face recognition with a mask? Setting up face recognition on mobile banking will increase the risk of bank card being stolen… Are these claims true or false?
Nowadays, face recognition technology is applied to more and more scenes, shopping, checking in, unlocking, traveling, and even getting toilet paper in some places cannot do without it. With the increasing popularity of face recognition systems, we have gradually become accustomed to the cameras “lurking” around us, but can this device really be installed by any organization?
”Although the current laws and regulations of our country do not prohibit the collection and use of personal information in the consumer field, relevant documents all emphasize that information collection must be’legitimate, proper and necessary’ and requires the consent of the parties; information use must be secured and not leaked When personal information is infringed, the infringing party shall bear the corresponding responsibility.” Yan Huaizhi, director of the Institute of Computer Network Attack and Defense, Beijing Institute of Technology, said in an interview with reporters, “In this sense, any organization When using a face recognition system to obtain personal information, it should comply with the principle of’legal, justified, and necessary’, rather than just install it if you want.”
”Face recognition technology is a universal technology that can be used in different Industry or field.” Yan Huaizhi believes that it is precisely because of this technical feature that many organizations use video surveillance as the name, first by setting up cameras to obtain facial information, and then using the system running in the background, without the permission of the parties. , Perform face recognition, and even further realize personal digital portraits in cyberspace through the identity information recognized by the face, which may constitute a violation of personal information and privacy.
Due to the needs of epidemic prevention and control, most people are now accustomed to wearing masks when they go out. However, there is a saying on the Internet that the mask must be removed to successfully pass face recognition. Is it really that troublesome?
”Wearing masks, hats, sunglasses and other items will indeed reduce the accuracy of face recognition, but technically it is not helpless.” Yan Huaizhi believes that from a technical point of view, a combination of global and local features of the face can be used. Perform face recognition, supplemented by three-dimensional data of unoccluded parts for three-dimensional face recognition.
In addition, it is reported that in order to adapt to the information management of epidemic prevention and control, face recognition thermometers have been used in many places on different occasions. Even if the subject wears a mask, the device can complete face recognition. For example, the intelligent anti-epidemic security gate set up at the 3rd China International Import Expo can use artificial intelligence algorithms to achieve temperature measurement and face recognition even when the monitored subject is wearing makeup, glasses, and masks.
Currently, many mobile banks support “face-swiping” transfers. Some people worry that this setting will increase the possibility of bank cards being stolen. Is this worry necessary?
“The application scenarios of face recognition usually include consumption and security. Consumer scenarios require extremely high accuracy of face recognition technology. The recognition rate of scenarios involving mobile banking “face-swiping” payment is usually above 99.99%, which is difficult for others. “Realize stealing.” Yan Huaizhi said, “At the same time, in addition to face recognition, in order to complete the transfer on mobile banking, mobile phone number, password and other information are also required. This information builds a safer protection network for online payment. Therefore, , The security of face recognition in mobile banking is very high, so users don’t need to worry too much.”
There are rumors that face recognition technology cannot recognize faces after plastic surgery. Is this really the case?
”Face recognition technology has reached industrialization and large-scale application standards. Common recognition problems such as simple plastic surgery and makeup have long been overcome, and the recognition capabilities of machine systems have surpassed that of humans. Therefore, whether recognition after plastic surgery can be achieved depends on the post-plastic surgery. The degree of change in facial features. If only cosmetic plastic surgery is performed, such as raising the bridge of the nose, shrinking the chin, or cutting double eyelids, the current technology can achieve accurate recognition; but if the accident causes a large area of disfigurement, face recognition is powerless Yes.” Yan Huaizhi said.
Take up legal weapons to protect faces that may be “lost” at any time. With the popularization of face recognition technology, like other personal information, people are beginning to worry about the leakage of face information. Because once it is leaked, it cannot be recovered. In the era of “brushing face”, how should we protect our “face”?
First, let’s understand the principles of face recognition technology.
Yan Huaizhi introduced that biometric technology, such as fingerprint recognition, iris recognition or face recognition, has been widely used in verifying the identity of the subject. Among them, face recognition technology has become one of the research and application hotspots in the field of computer vision and biometrics after half a century since it came out in the 1970s. The technical principle is to use a variety of measurement methods and technologies to scan human faces to identify or verify the identity of the subject in the image or video.
A practical face recognition system usually consists of modules such as face detection, face alignment, face characterization, and face matching. Traditional face recognition methods highly rely on artificially designed recognition features and machine learning technology. In recent years, the updated face recognition technology mainly uses deep learning methods based on convolutional neural networks.
Yan Huaizhi believes that because face recognition uses non-contact collection and non-sense recognition technology, it is very possible for individuals to collect facial information without noticing them. Personal biological data such as faces and fingerprints, like personal genetic data and health data, are special types of personal information that need special protection. In my country’s current laws and regulations, the Civil Code, the Cyber Security Law, the Personal Information Protection Law (Draft), the Data Security Law (Draft), etc. all provide protection for personal information including citizens’ facial information. Clearly stipulated.
Yan Huaizhi also said that the technology itself is neutral and innocent, but the use of technology should be within certain rules and scope. Protecting personal data such as faces and fingerprints requires the joint efforts of all parties, and personal efforts alone are far from enough.