"Face recognition" application scenarios continue to expand
"Face brushing" is more convenient and safer (Online China)
Our reporter Kang Pu
Face-swiping to unlock, face-to-pay, face-to-campus...In recent years, face recognition technology has continuously made breakthroughs, and the application scenarios have gradually expanded, which further facilitates our lives. It is even more effective in the normalization stage of epidemic prevention and control. At the same time, there are still some problems that plague the development of the industry, such as excessive introduction in unnecessary scenarios and the risk of information leakage.
Good helper for normal epidemic prevention
"Verification is successful, please pass!" Brushing his face, measuring the temperature, passing through the turnstile, and entering the station "seconds passed" without any contact yet-this was a scene at the Harbin subway station in Heilongjiang. Since July of this year, the "face-brushing" ride system has been officially put into use in Harbin Metro. After completing the mobile phone registration and facial recognition authentication, passengers can swipe their faces and ride the train without removing their masks. This is the first time in China to wear masks and swipe their faces take the tube. "During the epidemic prevention period, you can brush your face and pass without removing the mask, which is convenient and safe." said Ms. Liu, a citizen of Harbin.
The relevant person in charge of the operation department of the Harbin Metro Group said that using the above-mentioned face recognition system, passengers' entry and exit time, location and other information can be accessed in the background to achieve precise tracking and become an effective means of responding to the epidemic. It is reported that there are also urban subway stations in Shenzhen, Hefei, Wuxi and other cities that realize the "contactless" approach to the station to take the face temperature measurement.
Face recognition systems are also useful in epidemic prevention in scenic spots. After pre-face recognition registration and purchase of tickets, you can enter the scenic spot in a few seconds by raising your head, swiping your face, and passing quickly. At the same time, the temperature measurement is completed. Compared with traditional methods such as QR code verification and manual ticket checking, face recognition can quickly enter the park without more cooperation actions. This is very useful for reducing the time of queuing for ticket checking in popular scenic spots, alleviating the pressure of manual ticket checking, and improving the operational efficiency of scenic spots. It is helpful, but also conducive to park information collection and tracking.
As the epidemic prevention and control enters the normalization stage, the order of production and life is restored in an orderly manner. Traffic travel, entering scenic spots, going to theaters to watch performances, or entering office spaces and construction sites, rapid personnel verification and temperature measurement have become rigid requirements in crowded places. , The demand for "non-contact" has increased sharply. The "face recognition + temperature measurement" combined with face recognition technology and infrared thermal imaging technology can effectively reduce personnel contact and become a good helper for normalized epidemic prevention.
More competition on the track
For the widespread application of face recognition, the new crown pneumonia epidemic is only a catalyst. In recent years, the state has vigorously supported new infrastructure construction, and technologies such as 5G and artificial intelligence have accelerated the implementation of technologies to promote the construction of smart cities. Face recognition, as an important part of it, has developed rapidly. Face recognition for rides, face payment, face unlock, face recognition to enter campus, face recognition to register for marriage... With the advancement of technology, face recognition often appears in consumption and identity authentication scenarios.
Face recognition can improve recognition accuracy and reduce verification difficulty, which has great market potential. According to data from the Prospective Industry Research Institute, the market size of this industry reached 2.51 billion yuan in 2018, and the average annual compound growth rate of the industry market size from 2010 to 2018 was 30.7%. It is expected that the market size will exceed 10 billion yuan in 2024, which is accelerating. Enter the "Era of Visual Internet of Things". Internet giants, industry veteran brands and many up-and-coming stars compete fiercely on this track, dividing the tens of billions of "cakes." Hikvision, Dahua, etc. have already taken advantage of scale and have a solid market position, but there are also many entrepreneurial companies that have made breakthroughs in subdivisions and promoted industry technology upgrades. In 2019, the number of industry patent disclosures was as high as 6,700, while in 2009 this number was only 200.
In the past, face recognition often needed to recognize the entire face information. Chengdu Zhiyuanhui Company updated and optimized algorithms to achieve accurate recognition by extracting partial facial information. "The information from above the bridge of the nose to below the forehead is sufficient to help the system complete information comparison and identification verification." said the head of Chengdu Zhiyuanhui R&D.
The broad market prospect of face recognition is favored by capital. In 2019, the industry investment exceeded 6.5 billion yuan. It is expected that as the application scenarios continue to improve, more investors will enter the market, injecting greater vitality into the development of the industry.
Where to go for "human face"
While face recognition has brought convenience to epidemic prevention and control, production and life, some chaos has also caused widespread concern.
"My mobile phone's face recognition can occasionally be opened by others." It is reported that face recognition technology can be roughly divided into 2D face image technology and 3D face image technology. Some mobile phones use 2D face image technology with a lower threshold. Collect and compare portrait information, which can easily be verified by others through photos and other forms to crack the phone password.
"What should a business do if you resell your facial information?" "I don't know when someone will secretly ask you to pay by scanning your face when you pass by!" "Check your face at work, check your face when you eat, check your face when you stay in a hotel, or even go to the public toilet Wash your face with toilet paper.” Many netizens complained about overuse of facial information and easy leakage of information. As biometric information, human faces are unique and cannot be changed under normal circumstances. This is different from the disclosure of information such as mobile phone numbers and passwords, and it places stricter requirements on information collectors and managers.
Actions to rectify chaos in the industry are already on the way. The formulation of the national standard for face recognition will be fully launched in 2019. 27 companies including SenseTime, Tencent, Ant Financial, iFlytek, and Xiaomi have been included in the standard formulation working group. Companies have extensive experience in the entire chain of technology research, product development, application practice, and operation and maintenance guarantees. According to relevant companies, the working group will focus on personal data management specifications, technical specifications, and testing methods, and systematically build a national standard system in the field of face recognition to promote the healthy development of the industry.