“Brush face” payment when shopping, unlock “brush face” when using mobile phone, and “brush face” open door when entering the community Nowadays, more and more things can be “washed” and solved by face recognition technology.
According to a report released recently, more than 90% of the respondents have used face recognition, but 60% of the respondents think that face recognition technology is likely to be abused. Another 30% of the respondents said that they have suffered privacy or property loss due to the leakage and abuse of face information. So, is our face information safe in the “face brushing” era?
What are the risks of face recognition?
For face recognition, most people love and hate it. What they love is that it is convenient and fast, and what they hate is security risks.
Suspect “Ai face changing” deceives face recognition to commit crime
In the two cases of police embezzlement of citizens’ personal information this year, it is worth noting that the suspects used the “AI face changing technology” to illegally obtain citizens’ photos for preprocessing, and then generated dynamic video through “photo activation” software, deceived the mechanism of face verification, so that they could commit crimes.
Can masks replace faces to unlock mobile phones?
In the test, researchers put a mask on the opposite side of the mobile phone and then adjusted the light, color temperature and angle. After several comparisons, the phone was successfully unlocked.
Experts say the cost of making this mask is not high, and 3D printing technology can produce face masks or headgear with acceptable accuracy. As long as it is not in a very dark or bright background, the success rate of face recognition by mask or headgear is as high as 30%.
How to enhance the security of face recognition?
Experts say that the simplest face recognition at present only needs to collect and extract six or eight feature points on the face. For complex face recognition, it needs to collect and extract dozens or even hundreds of feature points on the face. Compared with unlocking the mobile phone, “face brushing” payment and “face brushing” into the community, the collected face feature points are more, and the security is naturally higher.
In addition, we have developed a biometric based in vivo detection technology, which can effectively identify the vital signs of the scanned object, and greatly reduce the risk of the recognition system using photos or masks as faces.
2 yuan can buy thousands of photos, how is the face information leaked?
According to the survey, on some online trading platforms, thousands of face photos can be bought for only 2 yuan, while the price of more than 5000 face photos is less than 10 yuan.
The store’s material library is full of personal privacy, such as real life photos and self photos. When the reporter asked the customer service whether the pictures involved copyright, the customer service denied it, but could not provide any material to prove the copyright of the photos.
If these face photos containing personal information fall into the hands of criminals, the photo owners may not only encounter accurate fraud and suffer property losses, but also be involved in criminal proceedings because their face information is used for money laundering and other illegal and criminal activities.
In the opinion of experts, the risk point of face recognition technology is more concentrated in the storage link. Due to the diversity of face recognition applications and the lack of a unified industry standard, a large number of face data are stored in the centralized database of application operators or technology providers.
Whether the data is desensitized, whether the security is in place, what is used for algorithm training, and which will be shared by partners is unknown to the outside world. Moreover, once the server is intruded, the highly sensitive face data will face the risk of leakage.
The draft of personal information law is issued, and the protection of face information is more standardized
In order to plug this loophole, experts put forward a variety of technical improvements, and further pointed out that face data storage should establish more stringent standards and norms, technology developers and app operators should collect, use and store data under more stringent supervision, laws and industry standards.
In view of the phenomenon that face information is abused, embezzled and collected at will, legal experts point out that the network security law explicitly includes personal biometric identification information into the scope of personal information. China’s “Civil Code” stipulates that the collection and processing of personal information of a natural person should follow the principles of legality, legitimacy and necessity, and obtain the consent of the natural person or his guardian, and the consent of the adoptee can also be withdrawn.
At present, the personal information protection law of the people’s Republic of China (Draft) is open to the public for comments. The draft proposes that the installation of image acquisition and personal identification equipment in public places should be necessary for the maintenance of public security, comply with the relevant national regulations, and set up prominent prompt signs. The collected personal images and personal identity information can only be used for the purpose of maintaining public security, and shall not be made public or provided to others, except with the individual’s individual consent or otherwise provided by laws and administrative regulations.