Biometric technology goes from laboratory to public

How old is he or she? I know it by tapping the phone. There was a small game that tested the age based on facial features, which burst the circle of friends. The reason why this game application can accurately predict age, gender and other information lies in the embedding of face recognition technology, and the function of this technology is far more than that.

Face recognition technology is a type of biometrics (hereinafter referred to as biometrics) technology. In recent years, biometrics technologies represented by face recognition and iris recognition have moved from the laboratory to the public, and upgrading people's "passwords" will make our lives more convenient and safer.

What is biometrics

Everyone has a unique iris texture, and it will remain the same for life

Biometrics is simply a way to use a person's physiology (fingerprint, face, etc.) or behavioral characteristics (sound, gait, etc.) to identify a person. Generally, the biological characteristics of the human body have the advantages of being carried around and difficult to forge, and have advantages in security, confidentiality, and convenient identification.

The common biometric identification method is fingerprint identification, but fingerprints are easy to leave traces to be copied by others, and have higher requirements on finger cleanliness and dryness. With the advent of the Internet of Everything, data security and information privacy are becoming more and more important, and people need more secure and effective "passwords" for identity authentication.

Face recognition is currently a relatively mature biometric technology. Because each person's facial features and facial features are different, this makes it possible to use these features to identify personal identities. Zhang Xin, expert member of the National Information Technology Standardization Committee Biometric Recognition Sub-Committee and marketing director of Beijing Kuangshi Technology Co., Ltd., said that the physiological characteristics of human faces can be converted into data information recognized by computers. The realization of face recognition technology is to first detect the face in the image or video, and then locate the key points of the target facial features, extract the features and perform structural processing. With these data, the computer can analyze the face attributes and understand the face.

"The improvement of face recognition technology benefits from the application and development of artificial intelligence related theories. At present, mainstream face recognition technology mainly combines deep learning algorithms. The more faces the machine" sees ", the more training data and deeper understanding , The more accurate and faster it judges the face, "Zhang Xin said.

Iris recognition is another type of biometric technology that is favored by the industry. The iris is a circular area located between the black pupil and the white sclera on the surface of the human eye. It contains many intertwined spots, filaments, crowns, and crypts. Sun Zhenan, a researcher at the Institute of Automation, Chinese Academy of Sciences, said that iris recognition technology confirms the identity of people by comparing the similarities between iris image features. The core step is to use pattern recognition, image processing and other methods to describe the iris features of the person to obtain data values. By matching and classifying the acquired iris data with the stored iris information, the purpose of confirming the identity of the person can be achieved.

He Zhaofeng, member of the International Biometrics Technical Standards Committee and deputy general manager of Zhongke Hongba, said that iris has many advantages as an identity recognition. First, iris recognition is highly secure. Everyone has a unique iris texture, even the same person's left and right eyes or identical twins, the iris texture is also significantly different. Secondly, the iris texture is formed from the third month to the eighth month of embryonic development, and remains unchanged for life, with good stability. In addition, the iris image is collected without contact, which is convenient and easy to use.

Like face recognition technology, iris recognition technology breakthrough also benefits from the development of artificial intelligence. Sun Zhenan said that the outbreak of biometric technology in recent years is not solely driven by the identification technology itself, it is an inevitable result of the development of social informatization to a certain extent. For example, the development of smart terminals has made iris data collection more convenient, and big data analysis methods have improved the accuracy of data comparison. The interconnection of all things and the subsequent higher requirements for security have made biometric technology Use it.

Lei Zhen, associate researcher of the Institute of Automation, Chinese Academy of Sciences, said that in recent years, recognition technology, including face recognition, has improved. First, it has benefited from the improvement of computing power brought by new chips and new algorithms. Second, smart terminal devices such as mobile phones. The popularization of the Internet makes large-scale and low-cost data collection possible.

What biometrics can do

Shopping in the mall, withdrawing money from the bank ... These daily business "brushing face" and "eye brushing" can be completed

At present, face recognition technology has entered the public's field of vision.

When I rushed to the office early in the morning and held hot coffee, would you worry about not being able to free your hand to swipe your ID card? At night, when you came back home from work and took out your pocket, would you be depressed because you forgot to bring the door key. With the face recognition system, all these worries will no longer exist. You only need to "swipe your face", the access control will automatically unlock, and you can even send the appropriate greetings according to your mood.

Intelligent security is an important application field of biometrics. Due to the characteristics of people with different biological characteristics and the same life-long, they can play a great role in security monitoring. For example, the use of face recognition technology, through dynamic monitoring, image search, portrait comparison, etc., can not only understand who has entered your community, but also help the public security organs to improve the security system.

Zhang Xin said that many fugitives have “bleached” themselves by pretending to be the identity of others, but face recognition can “see” the true identity of such people, so they can pull off cocoons in the massive population, accurately identify and report to the police.

In the field of intelligent business, the application of biometrics is gradually expanding. "To open an account at a bank, you must hold an ID card to verify your identity, which not only delays time, but also causes a lot of trouble for people with reduced mobility. With a face recognition system, you can complete real-name verification online." Zhang Xin said. In the future, when traditional industries such as finance, telecommunications, government services, and education move online, real-name verification will become more important. In these fields, there is a lot of imagination for the application of biometrics in the future.

The popularity of smartphones is an important driving force for biometrics. Instead of entering a password, "swipe face" and "swipe eye" can be used for shopping and payment. Zhi Liang, assistant general manager of Zhongke Hongba, said that some biometric technologies, including iris recognition, have become an important integrated part of smart phones, improving the security of mobile phones and realizing functions such as unlocking, file encryption, and payment security.

"Shopping in malls, going to the bank to withdraw money, taking high-speed trains to fly ... All kinds of daily business, 'swipe your face' and 'swipe your eyes' may be completed. It may even generate a large-scale biometric database, using biometrics, you Prove your identity on a global scale. "Zhi Liang said.

Experts also said that biometrics will not completely replace traditional identification technology, and some applications will be supplemented and supplemented by traditional methods such as passwords. Biometric technology and traditional models will complement each other, and people can use the most user-friendly, most efficient and safe identification method for different application scenarios.

Biometrics is here

China has reached the international advanced level in algorithm, and is in the forefront of the world in some applications

Lei Zhen said that as far as biometrics theory is concerned, China and developed countries are synchronized, and some algorithms and models have also received international recognition. However, developed countries such as the United States are in a leading position in new concepts and methods of artificial intelligence, and our country is more of a follow-up research. The earliest development of fingerprint recognition technology in China, face recognition, iris recognition and other popular biometric authentication began in the mid and late 1990s.

Take face recognition as an example. The team led by Defiance Technology has repeatedly ranked first in several world open algorithm competitions. The Zhongba Hongba team also won the championship in two consecutive international iris recognition algorithm competitions.

However, Zhang Xin believes that the importance of biometrics is the performance of practical applications. "In reality, the face recognition problems we encounter are far more complicated than the competition. We are more accurate in dealing with Asian face recognition, mainly because we have more data collected and trained in this regard."

He Zhaofeng said that Asians have dark brown irises and fewer iris patterns, which makes identification more difficult than European and American blue irises. At the same time, Chinese eyes are generally affected by eyelid occlusion, which causes the iris to not be fully exposed, and the more downward bending of the eyelashes further aggravates the occlusion, which puts higher requirements on the algorithm. He believes that in terms of algorithms alone, China has reached the world's leading level.

In order to cope with the development trend of biometric technology and huge application potential, major Internet companies have increased their layout in this area in recent years. Google has acquired a number of face recognition companies, Facebook has acquired Israeli face recognition technology companies, Microsoft, Samsung force mobile phone iris recognition technology. In China, Baidu, Tencent, and Alibaba have also begun to deploy in the field of biometrics. Industry experts predict that more companies will join in biometrics R & D and industrial promotion in the future.

Zhang Xin said that from the perspective of industrial application of face recognition, the United States and Israel are at the forefront of dynamic recognition, while China's application in many aspects is at the international advanced level.

Large-scale and effective data is the basis for developing applications, and unified standards are important conditions for the development of the industry. Zhang Xin said that developed countries started earlier in standardization, but China's current industry standards system is not yet complete and its coverage is limited. It should step up its organizational efforts to do standard-setting work.