I. Problems and Challenges
With the growing demand for personal identity authentication and management in the information society, biometric identification technologies and related products have entered a large number of aspects of social life, contributing to the continuous improvement of the quality of human life. However, there are some problems in the actual application process of biometric identification technology. At the same time, people have also raised questions and challenges for some biometric identification technologies that have been widely used. For example, human fingerprints can be easily duplicated and forged, and there is a possibility of using a forged fingerprint copy to deceive the fingerprinting system. Moreover, it is not difficult to obtain human fingerprints through certain technical means for counterfeiting. In 2006, the American popular science program MythBusters used a gel material that mimics the characteristics of human tissue to make a fingerprint copy of the human body, and then used this forged fingerprint copy to successfully pass the authentication of the fingerprint identification system. At Defcon16, the 16th global hacking conference held in 2008, ZacFranken also successfully certified the palm reader using dental artificial alginate materials and silicone rubber fake palms. In 2009, Duc Nguyen made it very easy to use a life-sized black-and-white picture to pass the user authentication of the face recognition system used in the Lenovo notebook.
Why is there such a problem? It starts with the principle of biometrics. The fact that biometrics (specifically the biological characteristics of the human body here) can be used as an effective means of identifying and identifying personal identities is determined by its own four characteristics: universality, uniqueness, stability, and non-replication. Sex. Universality refers to the fact that the characteristics of the human body on which biometric recognition depends should be innate for everyone, unless it is caused by special circumstances such as disability or birth defects. Uniqueness means that the biometrics are different from each other, so that individuals can be effectively distinguished. According to research and experience, each person's fingerprints, palm lines, iris, DNA, and subcutaneous vein structures are different from others. Stability means that the individual's biological characteristics can remain relatively stable within a certain period of time. For example, the human body's fingerprints, palm prints, irises, DNA, and vein structures are unchanged for the rest of their lives. Although faces and sounds vary with time, they remain largely unchanged for a specific period of time. Non-reproducibility refers to the difficulty in reproducing these biometrics due to the complexity and particularity of biometrics. From the above analysis, it can be seen that the universality and uniqueness of biometrics can be satisfied in most cases, while stability and non-replication are different due to the characteristics of various biometrics. Moreover, limited by the performance of sensors and biometrics algorithms, biometric systems will experience a reduction in recognition accuracy and anti-counterfeiting performance. For example, in theory, as long as the facial features of the human body are sufficiently large, even twins can be distinguished. In fact, for a realistic biometric system, it is almost impossible to do this. However, users do not need to be too pessimistic. People can adopt multi-biological feature recognition methods, ie, multi-modal recognition to improve system accuracy and ensure system security.
In addition to the above-mentioned precision and anti-counterfeiting, ease of use and user acceptance are also issues that need to be considered when using biometric systems. Taking the widely used fingerprint identification system as an example, if the user's finger surface is too moist or dry, it will greatly increase the possibility of the user being erroneously rejected. Moreover, the fingerprint identification system mainly adopts a contact acquisition method to obtain a human fingerprint, and the user needs to contact the surface of the sensor when in use. This kind of usage may not have too many problems in the application environment with strong personal characteristics, such as logging in to a personal computer. However, when used in a public environment, the subjective acceptance of some users may be reduced due to personal hygiene issues. For example, access control and attendance fingerprint identification systems in office environments. In recent years, non-contact biometric identification technology is receiving more and more attention. It provides an effective way to overcome personal hygiene and user acceptance problems brought about by contact biometrics.
In the networked society, the ultimate goal of biometric identification technology is that people do not need to carry any auxiliary identity items and knowledge, and they can use personal biometrics to personally authenticate themselves in a networked virtual society and the real world. Identification. For example, people can perform access control and attendance operations through a network-based physical access control system, access and modification of files through networked logical access control, and financial transactions through networked biometric identification. In order to meet the needs of a networked society, the gradual establishment of a networked biometric identification system will be an inevitable trend in the future of biometric identification technology, and will also have a broad market prospect.
Second, multi-modal biometric identification technology
The use of multi-modal or multi-biometric fusion techniques can achieve better identification performance and reliability than a single biometric identification system, and increase the difficulty and complexity of forgery of human biometric features and improve the safety of the system. The multi-modal biometric identification technology refers to a biometric identification technology that utilizes multiple identification technologies from the same biometric feature or multiple identification technologies from different biometric features to determine the personal identity. The development of multi-modal or multi-biometric fusion technologies is emerging as a new trend. A variety of multi-modal biometric identification technology solutions have emerged. For example, SecuriMetrics, Inc. of the United States designed a portable biometric recognition system that integrates fingerprint recognition, face recognition, and iris recognition—HIIDE, and applied it to the Iraq War and the Afghanistan War. The practical application results show that this multi-modal fusion identification system has higher security. At present, the system has been adopted by the U.S. military and applied to the U.S. military's military agencies and military bases around the world. In addition, BeiKe Wisdom Technology Co., Ltd. has also developed a biometric recognition system based on human palm veins and palm veins. The advantage of this system is that it captures the images of different optical channels at one time and obtains the palm of the human body. Patterns and subcutaneous vein images are identified. Compared with other multi-modal biometric identification products, this system has better usability and user acceptance, and it can complete multi-modal biometric identification with only one motion and high recognition accuracy.
From the technical realization point of view, multimodal biometrics recognition technology acquires different biological characteristics through various independent or multiple collection methods that are combined into one, such as fingerprints, palm lines, veins, and human faces. Image, iris image, etc. However, on the level of content analysis and decision-making, it can be subdivided into two implementations. One is to use different biometric algorithms to process different biometrics, calculate the matching values ​​of these biometrics, and then perform comprehensive analysis and judgment based on the obtained matching values ​​to obtain the final recognition result. For example, a fingerprint recognition algorithm may be used to analyze the fingerprint, a face recognition algorithm may be used to analyze the face image, and a final judgment may be made based on the calculated matching results. The other way is to integrate the collected various biological characteristics using a fusion algorithm, and to obtain a final determination result based on the comprehensive judgment of the multimodal fusion biometric identification technology. For example, the collected palmprint image and the palm image are fused, and then the image processing technology is used to analyze the content, and finally a comprehensive discrimination result is obtained. It can be seen that the first method requires separate processing of various biological features acquired, and can fully utilize existing algorithms for rapid integration, but the computational efficiency is not high. The second method requires re-design of the algorithm. It is difficult to develop, but the calculation efficiency is high.
In practical applications, it may be that some single biometric identification technologies have met the requirements of customers, for example, fingerprint recognition technology or face recognition technology. Moreover, from a cost-effective point of view, for some low-security application environments, a single biometric identification technology can be used as a solution. However, in situations where the user has ample funds and high security levels, multimodal biometrics should be used as a technical solution. Analyzing the reasons can be summarized as the following three points:
The first is that multimodal biometric identification technology has higher security and can effectively reduce the risk of illegal intruders performing the system. For example, a skilled criminal may be able to forge fake human body fingerprint information easily, but it is much more difficult to simultaneously forge fingerprints, palm prints, and irises. If the system is re-incorporated into the human subcutaneous vein information, then this multi-modal biometric identification system is theoretically impossible to break.
Second, multimodal biometric identification technology has higher recognition accuracy. At present, various single biometric identification technologies already have a high recognition accuracy, and to achieve a higher level of technology, it requires a huge cost. However, by adopting a variety of biometric identification technologies in a rational way, it is easy to achieve new levels of technology and play a multiplier role.
Finally, multi-modal biometric identification technology effectively solves the overall practicality of the system. As already discussed above, it is difficult for a single biometric identification technology to fully meet the requirements of universality, uniqueness, stability, and non-replicability. For example, with respect to the fingerprint identification technology relative to the user's hand incomplete, the iris recognition system is relative to the blind person, and the face recognition system burns relative to the twins and the face. These systems based on single biometric identification technology often affect the overall usability and safety of their systems due to certain special users. Obviously, multimodal biometrics will solve these problems.
At present, the main reason for influencing the further promotion of multimodal biometric identification systems is the main cost, so government departments and the military are its main users. However, with the addition of more suppliers, there will be more products with good price-performance ratios, and multi-modal biometrics technology will also move toward broader application areas and markets.
Third, non-contact biometric identification system
Compared with contact biometric systems, non-contact biometric systems can provide users with better user experience and health guarantees, thereby greatly improving the user acceptance of biometric identification technology. The non-contact biometric identification system means that the user can complete the collection, analysis and judgment of the human biometrics without using the system in direct contact with the system. For example, face recognition, iris recognition, and voice recognition are non-contact biometrics. Fingerprint recognition and palmprint recognition are typical contact biometrics. In addition, there are some biometric identification technologies that can be contact or non-contact, such as palmprint recognition technology and palm recognition technology. At present, non-contact biometrics technology is gradually attracting people's attention, and more and more products are entering the market. For example, Japan's Fujitsu's PalmSecure series of products through the non-contact acquisition of human palm vein user identification, Panasonic's ET-300 iris recognition system through a high-precision camera to capture the human iris, North Branch Wisdom Development company Whois The series also collects human palm veins and subcutaneous veins in a non-contact manner.
Panasonic's iris recognition system
Whois palm vein recognition system from Beike Wisdom
With the continuous progress and development of society, people pay more attention to public health and personal hygiene in daily life. Therefore, a biometric identification system used in a public environment needs to ensure that the device does not become a transmission route for bacteria and diseases. There is no doubt that non-contact biometric systems have a natural advantage in this regard. Moreover, the non-contact biometric system also reduces the possibility of intentionally reducing the safety of the system due to individual user considerations in terms of hygiene. For example, during the use of a fingerprint identification system, it was found that individual users use only fingertips for registration and identification in order to avoid large areas of finger contact with the device. Obviously, this use will greatly increase the possibility of misunderstanding and artificially reduce the overall security of the system. In addition, the non-contact biometric identification system also prevents users from leaving fingerprints during use, which can effectively prevent users from divulging personal biometric information and prevent professional criminals from replicating personal biometrics for criminal activities. The non-contact biometric identification system can greatly improve the safety performance of the system while improving the ease of use and user acceptance.
However, the non-contact identification method will pose great technical challenges for R&D personnel. First of all, due to the non-contact acquisition of the user's biometrics, on the one hand, it brings convenience to the user, and on the other hand it also increases the difficulty of the system to collect data. This problem does not exist for the contact biometric identification system. . For example, when the iris is collected, the user needs to be able to quickly and accurately locate the position of the human eye and the iris in the case where the user's height and position are different, and a high requirement is placed on the detection algorithm. For touch-type fingerprint recognition, there is no such problem. Each time the user places the finger in a designated area, the difficulty of positioning is reduced. Second, non-contact biometric systems need to overcome noise interference in a complex environmental context. Non-contact recognition means that the system sensor and the user to be identified need to maintain a certain physical distance, which inevitably introduces some background noise. For example, when a voice sensor acquires a user's voice, various other noises are also collected simultaneously. When the distance is long or the environment is noisy, the user's voice may be completely submerged in the noise. Similarly, face recognition and palmprint recognition are usually based on image analysis. The strength of light and the posture of the user will affect the recognition performance of the system.
Although non-contact biometrics have brought many technical challenges, various suppliers have invested a lot of resources in research and development in this area, and have also achieved remarkable results. A considerable number of products have already entered the industry. market. It is believed that more non-contact biometric identification products will emerge in the future, occupying a larger market share and becoming the mainstream of biometric identification technology products.
Fourth, the network of biometric identification system
The modern society is an information-based network structure, and information is transmitted through network nodes in various ways. The networked biometric identification system will bring a lot of convenience to people's lives. The ultimate goal of biometric identification technology is that people do not need to carry any auxiliary identity items and knowledge, and only use individual biometrics to physically control and control Logical access control. For example, replacing biometrics with passwords allows people to conduct financial transactions more securely through the Internet. Use biometrics to encrypt files and distribute and read documents. The use of biometrics technology for customer occupancy management, and then distribute the customer's biometric information to the designated room, completely avoid the trouble of loss of the card. Through the networked biometric system, the company conducts physical access control and access management for the company, enhances corporate image and management efficiency, and so on. These scenarios will become reality in the near future. The networked biometric identification system will bring about fundamental changes in the management of society and organizations, and social and organizational management will also undergo a qualitative leap.
Because the network of biometric identification systems will become an inevitable trend, governments, companies, and individuals need to be fully prepared for this change. First, the government must carry out legislation and standardization work. Personal biometrics are special information involving national security and personal privacy. Therefore, it is necessary for the law to restrict and protect the use and management of individual biometrics and prevent them from being abused by the state, business, or individual and infringe the rights of others. The establishment of biometrics standardization work can effectively integrate social resources and avoid resource waste and redundant construction. Secondly, companies that manufacture biometric identification products must develop technologies that can effectively protect individual biometric information and prevent biometric information from being destroyed, stolen and copied during transmission and preservation in the network. Finally, individuals should strengthen their awareness of self-protection and prevent the leakage of personal biometric information so that others can use the information for illegal activities.
V. Conclusion
As a sunrise industry in the 21st century, biometric identification has begun to enter every aspect of people's lives. Its development trend will be various networked, non-contact, multimodal biometric identification technologies. The development of these technologies will inevitably Bringing the development of biometrics to a new level and bringing people's lives into a new world.
Note: The pictures used in this article are from the Internet.
[1] Zhe Qiang, Zhang Dapeng, Wang Kuanquan. Palmprint recognition technology, Science Press, 2006.
[2] Biometric identification technology is not yet mature - easy to crack. http://
[3] New trends in biometrics: face recognition and multiple biometrics. http://
[4] Lin He. Biometrics - Eternal Identity Card. http://#007
[5] Arun Ross and Anil Jain. Multimodal Biometrics: An overview. in Proc. of 12th European Signal Processing Conference (EUSIPCO), Vienna, Austria, pp. 1221-1224, September 2004.
[6] James. Wayman, Anil. Jain, Davide Maltoni, Dario Maio. Biometric Systems: technology, design and performance evaluation. Springer-Verlag, 2005.
[7] David Zhang, Xiaoyuan Jing, Jian Yang. Biometric Image Discrimination Technologies, Idea Group Inc, 2006.
Note: The pictures used in this article are from the Internet.
[1] Zhe Qiang, Zhang Dapeng, Wang Kuanquan. Palmprint recognition technology, Science Press, 2006.
[2] Biometric identification technology is not yet mature - easy to crack. http://
[3] New trends in biometrics: face recognition and multiple biometrics. http://
[4] Lin He. Biometrics - Eternal Identity Card. http://#007
[5] Arun Ross and Anil Jain. Multimodal Biometrics: An overview. in Proc. of 12th European Signal Processing Conference (EUSIPCO), Vienna, Austria, pp. 1221-1224, September 2004.
[6] James. Wayman, Anil. Jain, Davide Maltoni, Dario Maio. Biometric Systems: technology, design and performance evaluation. Springer-Verlag, 2005.
[7] David Zhang, Xiaoyuan Jing, Jian Yang. Biometric Image Discrimination Technologies, Idea Group Inc, 2006.