The Security of Face Recognition Has Attracted Much Attention. How to Solve It_ Taigewang Technology

After long-term development and iterative evolution of technology, face recognition technology has been relatively mature and widely used in intelligent transportation, security, e-commerce and other fields. According to the 2018 global face recognition equipment Market Research Report, the market value of face recognition equipment in 2017 was US $1.07 billion, and it will reach US $7.11 billion in 2025, with a CAGR of 26.8%. Although the global face recognition market has grown significantly, it has also led to the discussion of privacy exposure in the use of face recognition all over the world. At present, the application of face recognition technology is in full swing, but it is also controversial. In March 2018, the Belgian government issued corresponding regulations to prohibit private use of face recognition or other biometric based video analysis cameras. In June 2018, the rekognition face recognition project of Amazon and the U.S. government caused public protest, and finally the Orlando police department was forced to abandon the relevant plan. Why is face recognition often criticized? What is the problem? The author's analysis finds that it mainly comes from two aspects. First, there are security vulnerabilities in the face data acquisition port (camera) and transmission network. Devices with simple or no password are easy to be attacked. Second, the attribution and confidentiality of face recognition information are not clear, which is still in the legal blank area. For example, now it has become the standard intelligent monitoring for streets, subway stations, airports, customs and other public places. These cameras can collect public identity and location information, and infer the changes, dynamic trends and user preferences of people flow in a certain place. But who keeps this information? How to keep it? Who can view it? There is no unified standard, but once these information is obtained by malicious attacks, causing personal property losses and even endangering public security, who is responsible? Face recognition encryption technology came into being for the above reasons. At present, many countries and institutions have developed anti face recognition technology to ensure that the collected face data is stolen and abused. A professor at the University of Toronto, Canada, has invented an algorithm that can dynamically destroy the face recognition system by light conversion of the image, so as to prevent the face recognition software from capturing the relevant face information. The principle is that AI driven filters look for specific facial features and change some pixels, so people can hardly see the difference. The test results show that the system can reduce the accuracy of face recognition from nearly 100% to 0.5%. D-id, an Israeli Tel Aviv network security technology company, has developed a data protection technology called de identification. Its main goal is to protect the data that has been used for identity authentication, and generate photos and videos that cannot be recognized by the algorithm while maintaining similarity with the real face, To protect personal privacy and identity information from malicious reading by facial recognition technology. As a big country in face recognition technology, China has also developed countermeasures. Kuangshi technology said that Kuangshi will desensitize the photos after collecting them, and only extract the photo features. Even if these features are stolen, they cannot be restored, and the process is irreversible. Conclusion: face recognition encryption technology mainly achieves the purpose of encryption by modifying several pixels of the image, which can alleviate people's concern about privacy disclosure to a certain extent. In the future, face recognition technology will be more widely and deeply applied. However, technology upgrading can not solve all problems, and corresponding laws, regulations and industrial production standards need to be established as soon as possible.