Fingerprint scanners are pioneers in biometrics. Multi-mode authentication is a new trend
Billions of smartphones and tablets are expected to process and collect multiple types of biometric input in 2017, including face recognition, sound mode, and iris scanning. However, the first development will be a fingerprint scanner. According to Deloitte's global forecast, by the end of 2017, the proportion of smartphones or biometric tablets using face, voice or iris recognition will not exceed 5%, while users using fingerprint readers will account for 40%. %12
Although the fingerprint reader is a sensor that has only recently been introduced, it has been all the way. Since the mobile phone is implanted into the microphone, voice recognition becomes a viable biometric input. For devices equipped with front-facing cameras, iris and face recognition are possible, but the quality of the lens and processor can affect speed and accuracy.
The challenge with voice recognition is that this technology can be problematic in noisy environments. Also, some people think that voice recognition can be distracting or unsocial, such as in an open-plan office or during a meal. Potential criminals can easily obtain sound information through recording. A viable solution is to combine voice recognition with challenge-response techniques, such as repeating specific phrases or answering security questions.
Face recognition typically requires similar lighting conditions as when the reference image was taken. If not, there may be cases of underreporting. 13. Even glasses, hats and scarves will affect the effect. Iris recognition requires the positioning and specific lighting conditions, and is sensitive to reflections, so wearing contact lenses or glasses will affect it. Another problem facing face recognition and iris recognition is that it is easy to encounter spoofing: Both are easily spoofed by face or eye photos. The alternative is to use interactive face recognition. For example, some systems require the target to blink. However, making clever videos or a series of photos can still fool the system.
In contrast, fingerprint recognition can be used in a dark environment and can be carried out while the user is walking, even if sitting in the back seat of a car driving at a high speed on a bumpy road.
From biometric phones to other devices
Biometrics like fingerprints are still a new attempt for smartphones, but it has been in use for decades. However, smartphones make everyone more familiar with the daily use of biometrics, while also eliminating some of the taboos associated with biometric data.
According to Deloitte's global forecast, the widespread adoption of smartphone biometrics will be a catalyst for deploying biometric sensors to other environments. For example, finger veins and palm vein scanners that use infrared to view venous tissue can be integrated into an automated teller machine, replacing PIN inputs; or they can be incorporated into the certification process for large transfers between companies. The school can use a vein scanner to confirm that the student has entered or left the building. Students can also use this technology to sign in and sign out in class or pay for food and school supplies. A growing number of countries may be considering the use of biometrics for national identity verification programs. As of now, India is huge, collecting human face, fingerprints and iris data. As of 2016, the number of registered people has exceeded one billion
Biometrics can be traced back to a thousand years ago, but it has been adopted on a large scale in modern technology or in recent years, and it may become more complicated and effective in 2017 and the following years.
Many private and public organizations should consider how to use a growing number of fingerprint readers in good ways and the huge crowds who are used to using them on mobile phones.
The challenge is to determine which applications can use fingerprint readers and other biometric inputs to provide fast and secure authentication.
Financial institutions: Deloitte's research found that 43% of adult mobile phone users in developed markets use their mobile phones to view their bank accounts17. Banks can learn how to better benefit from biometric systems in areas such as fraud detection, customer visits or opening new accounts, payment authentication, etc.
Retailer – E-Commerce: A one-touch settlement can be provided using a fingerprint reader, but the consumer must download the relevant application and also enter the necessary information, such as credit card data and billing address. Deloitte's research also found that most mobile phone users do not download more than 20 apps on their mobile phones. However, it is enough to make payments fast and secure, which is enough to attract users to download an application.
Retailer – Entity Business: The Store Payments application uses Near Field Communication (NFC) technology to complete payment authentication by placing a finger on the sensor and placing the phone close to the NFC recognizer. This way, you don't need to enter a PIN.
Enterprise users – read data. Biometrics can replace password input, authorize users to read emails, access corporate intranets, and other similar services. The reading and certification of the timesheet can be completed with a simple touch. Deloitte's research found that the current level of utilization of enterprise applications is very low. As a simple and secure method of authentication, the adoption of biometrics is expected to be promoted.
Media company – online subscription service. Vendors of music, news, television, or other content are behind the paywall, which can require users to use fingerprints to verify identity, thereby controlling the sharing of illegal user accounts and passwords. A single user account can be bound to a set of fingerprints. Fingerprints are far more difficult to share than passwords.
Enterprise users – physical security. Biometrics can be used to control access to the building without having to rely on access cards. Biometrics are different from access cards, do not need to be swiped, and will not fall at home.
Government services: Biometrics can be used to provide another way to access services, such as tax payments, medical records, and even electronic voting. This approach may attract young people to vote. Currently, this group has a high degree of ownership and use of smartphones, but their voting participation is below average.
The prediction here mainly focuses on the use of fingerprint readers, but all aspects of our daily life are inseparable from smartphones, which allows it to be combined with other unique data, such as input mode and positioning information. In our view, the combined use of various biometric inputs (ie multi-factor authentication) will become increasingly popular20. It can provide more reliable certification. For example, a banking application can use both fingerprint and voice recognition, obtain initial access rights through fingerprints, and then further verify with voice input.
The fingerprint scanner identification method is more secure than password recognition, and is not easy to be cracked, and can effectively solve the problem of input inconvenience caused by increasing password complexity for improving password security. Also, the unique performance of fingerprint recognition effectively avoids multiple passwords of multiple accounts, memory inconvenience, and improved unlocking efficiency.
Deloitte found in the "China Mobile Consumer Survey 2016" that the proportion of smartphones installed in China has reached 49%, of which 43% use fingerprint recognition as the authentication method. The use rate of smartphone fingerprint recognition technology is significantly higher than that. Developed countries. Fingerprint recognition experienced rapid growth in 2016, and smartphone manufacturers have adopted it as a standard for mobile phones. Fingerprint recognition has the advantage that traditional password recognition can't match. It is expected that in 2017, password recognition will become the primary authentication method for smartphones.
The fingerprint identification method is more secure than password recognition, but there are still some problems that need to be solved by the technology provider. First, improve the accuracy of fingerprint recognition. The security of fingerprint recognition is mainly realized by the uniqueness of the fingerprint, but the fingerprint can deceive the fingerprint identification device by complex manufacturing fake. Therefore, improving the accuracy of the fingerprint recognition device can effectively improve the security of fingerprint recognition. Secondly, combined with positioning technology to improve the security of fingerprint recognition. At the same time as fingerprint identification, it is determined whether there is a security problem by positioning. If the location of fingerprint identification changes, the user needs to perform further verification through the third method. Furthermore, improving the convenience of fingerprint recognition, and improving the performance of the device, it does not affect the fingerprint unlocking in a humid environment.
The scope of face recognition is further expanded
The recognition accuracy of artificial intelligence face recognition robot has reached more than 90%, and the accuracy of the face recognition technology algorithm has been continuously improved to promote the coverage of face recognition technology. Major banks also deployed face recognition technology in the financial sector, China Merchants Bank opened the face payment service, and the face recognition ATM has passed the acceptance. The medical field is mainly used in the field of card issuance, medical examination, etc., and the facial information is supported by the hospital information system supporting face recognition, and the patient's medical card is bound by the face recognition in the medical examination process. The occurrence of medical errors. The field of education is mainly used in the invigilation system to prevent the forgery of documents and the test of substitutes by comparing the face information of the candidates in the examination room and the marking office. At the same time, domestic smart cities continue to develop in-depth, the coverage of urban monitoring facilities is expanding, the definition is constantly increasing, the density of cameras is increasing, and the reduction of face data collection hinders the role of face recognition in the field of public safety.
Also, BAT technology-related exploration, it is expected that face recognition will be applied to more rich scenes. Alibaba cooperates with the outside world to develop face recognition technology to improve the security of mobile payment. Tencent products will gradually adopt face recognition technology. Baidu uses face recognition as a key research project for deep learning.
Iris recognition technology is currently one of the biometric technologies. Iris refers to the fabric-like loops in the pupil, containing many texture features similar to spots, filaments, wrinkles, crypts, etc., which are photographed by special optical imaging equipment, and using the image recognition software of the iris recognition system. A unique algorithm is used to extract and identify the desired feature points from the image and encode them. In the identity authentication, the extracted code and the code in the database are matched one by one to achieve the purpose of identity authentication. As some of Samsung's mobile phones are equipped with iris recognition technology, they are applied in international mainstream brand mobile phones.
Compared with the fingerprint recognition methods that are widely used at present, the error rate of iris recognition technology is much lower than that of biometrics such as fingerprints and faces, and the error rate is as low as one in 5 million. At the same time, the acquisition of clear iris texture needs to be coordinated by the user, so it is generally difficult to steal the iris image of others. Even photos, videos or even inactivated iris can not be matched by iris recognition. Also, the iris does not cause wear and tear and is unrecognizable. In terms of application technology, China's iris algorithm technology and hardware foundation have matured. As the mobile phone industry giants get involved in the field of iris recognition, the popularity of iris recognition will show an accelerated trend.
In addition to mobile phone manufacturing, the coverage of iris recognition technology is likely to expand further. Due to the wide coverage of mobile payment, the single transaction amount has increased, and the security of mobile payment has become the primary factor affecting the customer experience of payment products. The mobile payment platform is also actively moving into the field of iris recognition. Alipay's security architecture solution already supports iris recognition. And promote it. In the field of public security, China has not yet established a unified database, but in the populous countries such as India and Mexico, it has begun to promote the registration of the National Iris.
Speech recognition certification development is limited
Voice recognition does not need to rely on expensive equipment, only the microphone can collect data, the user can speak a specific phrase through the microphone, through filtering noise, extracting features, feature comparison, reaching a certain approximation threshold to pass the authentication. Although voice recognition is low in cost, it has not been widely applied in the field of identity authentication due to security issues. The voice features available for extraction are significantly less than the fingerprints and faces, and the speech recognition is more susceptible to background noise and the authentication fails. Poor stability, a voice will be affected by age, health and other aspects, common diseases such as colds are enough to cause certification barriers. Also, simple recording or imitation can deceive the speech recognition system, posing a greater security risk. Therefore, it is vulnerable to external influences, has poor stability, and is easily imitated into three obstacles that hinder speech recognition technology. In the future, the prospects for the safety certification market will be limited.
Multimodal authentication becomes a new trend in identity authentication
Multimodal biometrics technology refers to identity authentication by combining multiple biometrics to improve authentication and accuracy. Each biometric technology has unique advantages and limitations. A single technology can effectively solve the identity authentication problem in a single scenario. However, due to the increasingly complicated application scenarios, a single authentication technology cannot meet the requirements. For example, face recognition and multi-factor verification such as eye pattern, the new multi-modal biometric identification method has an accuracy rate of 99.99%, which exceeds the single face recognition technology, and has been applied to the field of mobile payment. As multi-modality recognition technology becomes more mature and costs gradually decrease, its service scope will be further expanded, such as identity authentication in transportation fields such as airports and railway stations, identity authentication of cardholders in medical and social security fields, and access control facilities in residential areas.Fingerprint scanners are pioneers in biometrics. Multi-mode authentication is a new trend。