At the World Internet Conference last year, the "face-swiping" technology was a hit. Every Internet conference puts security work at an important position, this year's Internet conference is different from the past in terms of entrance and exit management, just swipe your face here. When passing through the security check channel, the photos and information are displayed on the big screen, so that the security check personnel do not need to go through the inquiry and go directly to the scene, which greatly increases the efficiency of customs clearance. It seems that biometrics has developed greatly in the field of access control.

The difference between RFID access control and biometric access control

Now access control has more advanced technologies and new application markets, and the technologies currently applied to access control systems include: barcode, magnetic barcode, radio frequency identification, biometric identification, etc. Among them, radio frequency identification access control and biometric access control are two major trends in the development of access control.

The biggest difference between radio frequency identification access control and biometric access control is the different authentication media and methods, radio frequency identification access control uses ID cards and Mifare cards as media, the authentication media is easy to lose and damage, and ID cards are easier to copy. Biometric access control currently has fingerprint, palm shape, face image, iris and other identification methods, the authentication medium will not be lost, and the security is better. Radio frequency identification access control and biometric access control have been widely adopted in different requirements and applications. However, due to the bottleneck of technological development, the biometric authentication model still has certain problems and difficulties in practical application in terms of stability, application cost, and difficulty in system construction. Of course, with the continuous development of technology, some of these problems have been solved. In order to improve the security and applicability of the system, multiple authentication modes such as smart card multiple authentication, smart card + password authentication, and biometric + smart card authentication have been developed.

Application of biometrics in the field of access control

The most widely used biometric access control systems include fingerprint recognition access control, face recognition access control, palmprint recognition access control and iris recognition access control.

Face recognition system

Face recognition access control is a face recognition access control and attendance product that can be run offline, it is positioned in the mid-to-high-end access control and attendance market, partially replacing the current swipe card and fingerprint access control attendance machines on the market.

The Internet conference is a time for companies to show their own technology, according to Baidu’s face-scanning access control, users only need to scan their Identification card and enter a real-time photo when entering the park for the first time, it takes within 10 seconds, after that, you can "swipe your face" to pass through the gates of Wuzhen Scenic Area. Baidu describes the application scenarios of this technology like this: "When the future cities apply the technology of face gates more to life, people may no longer need keys to get in the car and go home, and they may even use their faces to adjust household appliances such as air conditioners and washing machines.

In the internal letter, Baidu mentioned that the face recognition gate opportunity is based on the deep neural network machine learning algorithm to learn and recognize multiple facial feature points detected, the entire project runs through R&D, hardware, technical support, and deep learning laboratories, AI platform department, process information management department and internal communication department. Baidu's face recognition system has an accuracy rate of 98%.

Fingerprint recognition system

The fingerprint access control system replaces the traditional key with a finger, when using it, you only need to place your finger on the collection window of the fingerprint collector to complete the unlocking task, the operation is very simple and avoids other access control systems (passwords, identification cards, etc.). May be forged, misappropriated, forgotten, deciphered, etc.

The fingerprint recognition access control system uses fingerprint recognition technology to verify identity. Fingerprints are carried with you, everyone is different, and they will remain the same for life. RFID cards may be borrowed, but fingerprints cannot be borrowed. The fingerprint recognition access control system is more secure and accurate, and the media used will not be forgotten or lost. At present, the cost of fingerprint recognition access control system is equivalent to that of RFID card access control system, because fingerprints are free of cost.

Iris recognition access control

Iris recognition technology is currently the most accurate method of identification in biometrics, compared with other identification technologies, iris recognition has the following characteristics:

1.Accuracy: Commissioned by the British government, the National Physical Laboratory (NPL) tested and compared seven technologies including retina, iris, fingerprint, palm print, face, voice, and handwriting dynamics. According to the report, the accuracy of the iris and retina is the highest, and the face is "the least accurate." to improve the accuracy of fingerprints, ten fingerprints must be collected. In addition, the Japanese Automatic Identification Seminar (AIM) gave the false acceptance rate of different technologies, the iris is 1,200 times more accurate than fingerprints, 12,000 times more accurate than faces, and 40,000 times more accurate than voices. AIM believes that the least accurate of the seven technologies is voice recognition.

2.Anti-deception: NPL believes that the anti-spoofing iris and retina are the strongest, fingerprints and palm prints are easy to forge, signatures can be imitated, voices can be replaced by recordings, and faces are "the easiest to be deceived." For example, fingerprints will leave a printed film every time they are used, which can be easily obtained by others to make fake fingerprints.

3.Practicality: NPL believes that the retina is in the fundus, it is difficult to obtain images, and the development prospect is the worst. Dark iris is difficult to collect usable images (in fact, the collected images are difficult to correctly recognize). The sound requires a high-fidelity microphone. High-precision cameras are required for face and palm prints. Handwriting requires a special writing board, and cannot use the existing configuration on the computer. Fingerprint acquisition is easy, but it belongs to contact collection. The disadvantage of contact collection is that it easily contaminates devices and affects accuracy. In addition, for fingerprint recognition, there are problems such as physical labor causing texture wear, and dry fingers of the elderly that affect use. The iris can be obtained with a normal camera, which is harmless and convenient to the human body.

4.Security Level: Compared with face recognition, voice recognition, fingerprint recognition and palmprint recognition in the field of pattern recognition, iris recognition has a higher security level, but the sensor signal collection has special requirements for the operation of the user, which may easily cause damage, collectors are nervous and uneasy, so iris recognition is currently mostly used in fields and departments with higher levels of information security.

The application of biometrics in the access control system is already what many access control companies are currently doing, with the increase of users of biometrics products, users' acceptance of the use of biometrics products is also increasing. Different product applications will cause users to different levels of acceptance of biometric technology. Biometrics technology has begun to reach the level of large-scale application, including product price, quality and technology are relatively mature.