Home > News > Industry News > A wide variety of biometric me.....
Browse Categories
New Arrival(26)
Hot Products(10)
Reader & Modules Series(18)
Handheld Terminal Series(14)
POS Terminal(5)
Time Attendance Series(12)
Access Control Series(5)
Sales Promotion(35)
Door Lock Series(11)
Camera Series(11)
Explosion Models
Contact us
Tel:+86-23-67305242 :
Mobile:+86-13667681778
Email:info@hfcctv.com
Web:http://www.hfteco.com/
Facebook:Huifan Technology
Contact Now

News

A wide variety of biometric methods Ten billion market prospects can be expected |HFSecurity

  • Author:HFSecurity
  • Release on :2019-11-11
According to foreign media, this week, Michigan State University and New York University researchers jointly released a research report on fingerprint recognition: using some common features of human fingerprints to create fake fingerprints can easily deceive fingerprint sensors of smart phones.




On today's smart devices, the fingerprint recognition module is almost standard. People are increasingly dependent on it. Various applications such as unlocking and payment can be identified by fingerprints. In addition to fingerprint recognition, various biometric technologies such as face recognition and iris recognition emerge in an endless stream.

In many people's minds, biometrics have unique qualities that should be theoretically very safe, but in terms of technology and practical applications, biometrics are not a problem. Of course, this does not affect biometrics as the mainstream trend in the future, and does not prevent it from becoming a tens of billions of markets. Driven by the dual exploration of spirit and interests, human beings will eventually find a reliable way of biometric identification to promote the further development of the industry.

First, there are many kinds of biometric identification methods, but the theory is feasible but the practice still has defects.

Biometrics, which are closely integrated with high-tech means such as optics, acoustics, biosensors, and biostatistics, use the inherent physiological characteristics of the human body (such as fingerprints, faces, irises, etc.) and behavioral features (such as handwriting, sound, Gait, etc.) to identify individuals.

Before the emergence of a variety of biometric technologies, we generally identify individual identities by identifying the identity of the item, document or username, password and other identification knowledge. However, once such foreign objects are stolen, their identity can easily be used or even replaced. With the development of technology, biometric identification has become a new type of identification because of its uniqueness and non-replicability.


android handheld device with live fingerprint scanner



fingerprint scanner demo

At present, the main biometric methods are touch (fingerprint recognition, palmprint recognition, finger vein recognition, palm vein recognition) and non-touch (face recognition, iris recognition, eye pattern recognition, DNA recognition, behavioral gait recognition). ). Different methods of identification have their own advantages in theory, but there are still many problems in the process of practice.

1. Fingerprint recognition: the most common way of biometric identification, but security risks cannot be ignored

As far as the current situation is concerned, fingerprint recognition is the simplest, most accurate, and best-developed form of biometric identification.

Traditional fingerprint recognition methods are punching and access control, but with the popularity of mobile Internet, mobile payment, information security and other fields have a demand for biometrics. The earliest and most widely used fingerprint recognition has gradually spread to the mobile side.

Throughout the global biometrics market, fingerprint recognition accounts for nearly 60% of the total, with thousands of manufacturers and hundreds of products. Other biometric technology manufacturers are less than one-tenth of fingerprint recognition. Undoubtedly, fingerprint recognition has become the most recognized and accepted way for mainstream smart device manufacturers and consumers. On the eve of the outbreak of mobile payment, fingerprint recognition based on smart terminals has also ushered in a large-scale outbreak.

However, in practical applications, fingerprint recognition exposes a large number of problems. For example, it is easy to forge, contact type is not healthy enough, weather changes affect stability, etc., and its safety is also questioned. The above report on fingerprint identification is a deep reflection of this hidden danger. The researchers simulated a series of artificial "master fingerprints" by computer, and found that the matching ratio between the artificial fingerprint and the real fingerprint in the sensor reached 65%.

Although this method has not been tested on real mobile phones, information security experts believe that the probability will drop in practice, but this research can still lead us to think about the reliability of fingerprint recognition. Although each person's fingerprint is unique, the size of the sensor on the mobile phone is very small, and only part of the fingerprint can be scanned. After the user enables fingerprint recognition, the mobile phone usually obtains 8 to 10 fingerprint images for matching, not to mention many. The user has recorded more than one fingerprint. Once this information is leaked, it will create a huge security risk.

The scientific and technological circles have different opinions on the results of this research. Some believe that research has limitations that are not sufficient to prove anything, but more opinions agree that this result does reflect the risks of fingerprinting, as Andy Adler, professor of systems and computer engineering at Carleton University in Canada, said: "This is not the case. It’s worrying, but it’s really bad. If what I want to do is take your phone and use your Apple Pay to buy something, if I can crack a 1/10 phone, then the situation is very serious.”

2. Face recognition: The theory is good for commercial use, but there is still a distance from the actual application scene.

Face recognition has experienced considerable development time since its appearance. This method is non-contact from entry to identification. The time is short and the accuracy is guaranteed. It has been gradually accepted in China, and the application of boot unlocking, registration, payment, file encryption, etc. on the mobile side is gradually strengthened. Increase. In addition, there are also popularizations in the fields of family planning social security, judicial public security, housing construction, and education.

In China, there are not a few companies that have deep face recognition. In BAT, Ali defined the DT (data technology) strategy in 2104 and acquired face++ in 2015. Tencent has established a "Excellent Team" dedicated to image processing and pattern recognition. Baidu set up a deep learning team when Wu Enda left, and scored the highest score on the LFW database, which is known as the "most difficult face image library." Other companies such as 360 hired Yan Shuicheng, an associate professor at the National University of Singapore, as the dean of the Institute of Artificial Intelligence. Companies such as Sichuan Zhisheng, Hanwang Technology, and Green Squat are also involved in face recognition.

It can be seen that face recognition has reached a relatively frequent level at the commercial level. However, deep digging techniques and practical applications will find many defects.

On the technical level, the data on the LFW is actually not very good. For example, the accuracy of face recognition reaches 99.7% on LFW, which seems to be extremely high. However, when 99.7% of the technology is verified in the actual scene, it will find that its accuracy may be only 75%. . The real situation is that many of the pictures in LFW are downloaded from the Internet. The quality of faces is very different. Some people think that this is closer to reality, but it is still far away from most application scenarios.

biometric fingerprint scanner wireless




In the LFW, the brush is the contrast between the face and the face, namely 1:1, 1:N and N:N. However, in practical applications, all manufacturers have to reduce the N sample, or add other methods to ensure convenience and security. For example, if we go out to eat, we will face the camera and then leave when we check out. This is verifying 1:N. After half an hour, the server has not found the user's face. After all, it is necessary to search in hundreds of millions of databases, and the user experience can be imagined. But if you enter the user's name or other information at the same time, then this N is much reduced, switching from 1:N to 1:1, which improves efficiency.


The problem with this is that if someone attacks the user's account, the risk increases. There are two main conditions for current face recognition, one is identity and the other is living. Is it successful if you take someone else's photo and scan it for the camera and then sign the ID number? Based on living conditions, face recognition requires several expression transformations to improve security, but what if the victim's video is captured? This is also a question worth pondering.

3. Iris recognition: increasing accuracy, but need to cooperate and train

Iris recognition is a new type of recognition technology, because the human iris tissue will not change throughout life, and it has higher stability than fingerprints and appearances, so the recognition rate is also high, and its non-contact recognition method is also more easy. accept. However, due to the complexity of the technology, it is necessary to be able to identify smoothly under a certain angle and avoid direct illumination, and the action range is large. Therefore, it is necessary for the early users to have an active cooperation mechanism, that is, the process of receiving training.

However, biometrics are favored. In addition to security, efficiency is also an important factor. If you spend too much time on identification, it is better to use fingerprints or face recognition directly.

In summary, different biometric methods have their own advantages, but because the technology and industry are not yet mature, the disadvantages also exist. Even in such a situation, many companies can still see the succession of the company, in addition to the pursuit of new technology and innovation, but also because this cake is really attractive. It can be expected that the biometric market will be more popular in the future and will be able to reverse the further development of technology.

Second, the tens of billions of market to be developed, the prospect of biometrics can be expected

According to the "China Biometrics Industry Market Prospects and Investment Strategic Planning Analysis Report" published by the Prospective Industry Research Institute, the global biometrics market size was 3.422 billion US dollars in 2009, and the scale was almost 9.8 billion US dollars in 2013. Breaking through the $10 billion mark. The global biometrics market is expected to reach $25 billion by 2020.
According to the "China Biometrics Industry Market Analysis and 2015-2020 Development Strategy Research Report", the average growth rate of the domestic biometric market is more than 60% from 2010 to 2014. Year 2014. The biometric market is 8 billion yuan. It is estimated that by 2020, the biometric market will exceed 30 billion yuan.

Such a tempting cake, it is no wonder that related companies have entered the market, the tens of billions of markets, who do not want to take a slice of it? Of course, profits are not so easy to obtain. At the 315 party, a static photo was easily cracked by APP image recognition and dynamic synthesis technology.

bluetooth biometric fingerprint reader



Security risks such as these are not rare in the case of biometrics. Although biometric identification is an inevitable trend in the development of science and technology, the different vulnerabilities in different identification methods do hinder the development of this industry to some extent.

At present, many companies have developed their own algorithms for different biometric methods. The next step should be to achieve some breakthroughs in module acquisition and application. In the past, fingerprint recognition developed rapidly, and face recognition is now emerging. In the future, iris recognition, voiceprint recognition, and vein recognition are all areas that can be deepened and explored, and are also hotspot technologies that can be combined with each other.


In fact, the real development trend of biometrics technology should be to enter the life of ordinary people and meet the recognition needs of people in various scenarios. However, only one type of biometric technology can never meet all the needs of the scene. A high security level environment will inevitably require multiple biometrics to complement each other. The combination of fingerprints, face, iris, gait and other identification methods is a good development path.

The combination of biometrics is a reference for the relevant technical departments and enterprises. After all, there are still 10 billion markets waiting for development in the field of biometrics, and it is conducive to improving accuracy and safety, and may further promote the industry.

website:www.hfteco.com

Facebook:Huifan Technology

Youtube: Huifan Technology