What about the security issues of smart door locks? 32 out of 38 samples were unlocked by fake fingerprints?

The implementation of fingerprint recognition technology is considered one of the most disruptive innovations of the 21st century. 

Safe, accurate and fast, these are the great praises that fingerprint recognition has received after encountering smart door locks and smart phones; but the circulation of technology around the world has always been dialectical. It brings surprises to the world while also bringing some troubles.

Today, a Weibo post about a smart fingerprint lock being "cracked" continued to heat up.

According to the test results of online-sold smart door locks jointly announced by the Consumer Associations of Beijing, Tianjin and Hebei, among the 38 samples tested, 6 samples could not be unlocked in extremely low temperature environments; 8 samples had abnormal reactions in the electromagnetic compatibility test; the IC card keys of 24 samples could be cracked and copied; 30 samples had poor reliability; and 32 samples could be unlocked with fabricated fake fingerprints.

Not long ago, a video of using "orange peel" to unlock the fingerprint recognition of a mobile phone went viral on the Internet. By sticking a layer of transparent tape on a smartphone equipped with fingerprint verification function, the unlocking function of the mobile phone can be successfully "cracked".

Regarding fingerprint products that are frequently cracked, an industry insider explained that semiconductor sensors obtain fingerprint pattern information by relying on the voltage difference in the valley and ridge part of the fingerprint, which can also be understood as treating the fingerprint as a conductor.

"Fake fingerprints" are similar to a conductive coating. When a user touches the conductive coating to unlock the phone, the information received by the fingerprint sensor is actually the conductive coating instead of the fingerprint. After repeated use, the conductive coating will form a fixed pattern of part of the user's fingerprint, which is at risk of being copied by others.

In addition to fingerprint recognition, other options are available. In response
to some technical loopholes in fingerprint recognition, Peng Cheng, product technology director of Yanshen Technology, said that in the later stages of the development of fingerprint recognition technology, if we want to achieve some breakthroughs in security, we must have a qualitative breakthrough in fingerprint sensor related technologies. At the same time, we can also integrate more high-quality liveness detection and other technologies to achieve both surface features and hidden features.

In addition, he also mentioned that any technical system has loopholes. If users consider the security of smart door locks more, iris recognition is a better option.

In his opinion, fingerprint recognition technology is mature, but it is greatly affected by the environment in actual applications. The recognition rate of more than 99% in laboratory environment is difficult to achieve in actual applications.

The advantages of iris recognition are relatively obvious. According to relevant data, the error rate of iris recognition of one eye can easily reach one in a million or even lower; in addition, iris recognition is a non-contact recognition, which is very convenient and efficient.

In specific applications, the iris can extract 173 independent feature codes (also known as degrees of freedom), that is, the probability that two irises are exactly the same is one in 2 to the 173th power; while other biometric technologies have only 13-60 independent features.

Compared with face, fingerprint and other information, iris information is richer, containing many detailed features such as intersecting spots, filaments, crowns, stripes, crypts, etc.

After the iris is formed during the fetal development stage, it will remain unchanged throughout the life course. These characteristics determine the uniqueness of the iris characteristics and also determine the uniqueness of identity recognition.

In addition, Dr. Wang Dong, CTO of Beijing Smart Eye, who specializes in the research and development of biometric technology, said that compared with traditional fingerprint recognition, finger vein recognition may also be a better choice.

Finger vein recognition is a type of biometric technology. It uses a finger vein recognition instrument to obtain an individual's finger vein distribution map, and then extracts feature values ​​from the finger vein distribution map based on a dedicated comparison algorithm.

Finger veins are one of the characteristics of the human body. The hemoglobin in the red blood cells in the human veins is the reduced hemoglobin of oxygen. The reduced hemoglobin absorbs near-infrared rays. Therefore, when near-infrared rays are irradiated to the fingers, only the vein part will have a weak reflection, thus forming a vein pattern image.

The subsequent steps are similar to other biometric identification processes. After the vein feature values ​​are extracted and stored, advanced filtering, image binarization, and refinement methods are used to extract features from the digital image. A complex matching algorithm is used to compare and match the vein feature values ​​stored in the host, thereby authenticating and confirming the individual's identity.

 "In a sense, this technology is a kind of liveness recognition technology, which mainly relies on infrared light to illuminate the finger to obtain blood vessel patterns, and is a kind of liveness code formed by blood flow.

Compared with traditional fingerprint recognition technology, finger vein recognition is not affected by factors such as whether the finger is wet, clean or damaged (no contact with the instrument is required); it is also safer, faster and more accurate. The probability of two random people having the same finger vein structure is one in 3.4 billion, so it can be said that everyone's finger vein structure is unique; most importantly, the entire recognition process takes less than one second to complete. "Wang Dong said.

Smart locks, big business
today, the integration of intelligence has made traditional locks no longer just defense products, and the penetration of scientific and technological blood has given them unprecedented social value.

According to statistics, there are currently 1,129 smart lock companies and more than 2,800 brands in the country. The shipments in 2018 were between 15 million and 20 million sets, and the output value of the entire industry exceeded 10 billion yuan.

The popularity of smart locks instantly filled the track with many players.

Among them, there are old manufacturers that have been deeply involved in the traditional lock industry, Internet companies that are familiar with user experience, many well-known home appliance companies, and a large number of security manufacturers.

Today we are going to talk about the fourth type of players.

Security manufacturers have always been engaged in business targeting governments and enterprises’ safe city projects based on video surveillance.

In the past two years, Leifeng.com has found that whether it is Haidayu, the leader in the current security industry, or small and medium-sized enterprises such as Smart Eye and Eye Technology that are deeply engaged in security segmented scenarios, they have all launched smart lock products based on their own characteristics and technological advantages.

"To be precise, smart locks are the best fusion point of security video surveillance technology and Internet of Things technology, and they directly take on the important task of connecting the home Internet of Things.

In the future, smart locks will be like smart video surveillance. They are not dumb iron boxes, but smart nodes with vivid life. Every time a user enters or leaves the door, they are interacting with information. "

Zhou Jun, founder and CEO of EyeSight Technology, directly explained the innovative spirit of this product.

"In the future, it will not only play the role of a lock, but also be connected to the smart home backend platform to enable remote calls, remote unlocking, smart alarms, smart big data analysis, etc. It is one of the most core entrances for major security manufacturers to enter the C-end market, leverage end users, and explore IoT layout."

In 2015, some cities issued new regulations: public housing access control and keys must be linked to tenant information. The "Specifications" require that residents of public housing have permission to use access control and keys. 

According to statistics, there are about 460 million households in my country. In densely populated towns, about 20% of the population mainly rent houses, and the proportion of rental houses in first-tier cities is as high as 40%. From the perspective of urban development and urban resources, the rental proportion is expected to exceed 50% in the future.

In view of this special situation, the standardization of public rental housing management needs to be strengthened urgently. Traditional mechanical keys or access cards not only fail to bring convenience to users, but also fail to achieve standardized management of tenant information, causing serious security problems due to the flow of personnel.

At present, the security business is roughly divided into "Skynet" and "Ground Network". The client of Skynet is the public security, while the client of Ground Network is the community. In the Skynet project where traditional manufacturers are competing fiercely, the strong demand for Ground Network has also attracted many participants to join the race.

In the smart community system, the ground network project mainly collects facial information. However, due to reasons such as light and obstruction in actual applications, the relevant video probes often find it difficult to capture high-quality and comprehensive facial information.

As a component unit of the community, houses equipped with smart locks can collect almost all fingerprint data of people entering and leaving the community, except for faces.