Is the most popular face recognition that can be s

2022-08-23
  • Detail

Is the ubiquitous face recognition just a need or a gimmick

electromechanical; Electromechanical industry; Intelligent manufacturing; Electromechanical technology; Face recognition; Industry information in recent years, face recognition technology has been applied in more and more scenes. Whether it is the early face brushing payment or the face brushing car ride, now there are also fresh attempts to brush the face to take toilet paper, brush the face to throw garbage, and even brush the face to buy a Maotai, which has a rapid trend of penetrating all walks of life

contrary to the continuous laws and regulations issued by various states in the United States to prohibit face recognition technology, China is very positive about the promotion and implementation of this technology, and the loss risk of enterprises is relatively high. The most important thing for the generation and application of new technologies is to solve the current pain points and benefit the people's livelihood. Under the condition that the current verification method can still meet the needs well, what is the purpose of fully promoting the application of face recognition technology? Face recognition technology has been criticized repeatedly. How can we protect our privacy and security

what pain points does face recognition solve as the scene expands

face recognition has become a trend. Various fragmentary reports on face recognition have long been common, some are still being steadily promoted, and some have already quietly ended. However, when discussing whether a technology has the ability to achieve large-scale applications, we must think from the two dimensions of technology and business model

first of all, no matter how packaged in business, face recognition technology is not a cutting-edge technology. Large foreign technology companies, such as Amazon and Google, have been using it for a long time. Although the domestic application volume is large, the level of China is still not higher than that of the United States in terms of the basic research of artificial intelligence behind face recognition technology

however, China's commercial pace is many times faster than that of the United States. According to the data of research institutions, in 2016 alone, the market size of China's image surveillance system, including equipment and video management software, reached 6.4 billion US dollars. The government and private enterprises have installed 176million surveillance cameras, ranking first in the world. In the next five years, the compound growth rate of this business in China will reach 12.4%, and the application range of face recognition will only be wider

based on the analysis of the application scenarios of face brushing entering the station and face brushing gate, the application of face recognition technology does not solve the pain points of speed and accuracy. Since this year, Shenzhen, Jinan and other cities have successively opened this service, and Beijing is also planning to promote it. At present, this has also become a landmark application of urban intelligence, which is sweeping more cities

Taking Jinan as an example, the set value of Jinan "time interval" is the sampling interval when outputting text experimental data. South Metro Line 1 is the first subway line in China to adopt 3D face recognition gate, and there is a gate at each entrance and exit to realize face brushing. The head of the material management department of Jinan rail transit group said: this device comprehensively applies cutting-edge science and technology such as mobile payment and biometrics, and extracts face information and enters it through big data collection. Passengers can completely get rid of ID cards, boarding cards and other media. After successful registration, they do not need to carry any cards

but after a series of downloading, registration, login, information entry, payment binding and other processes, it can finally pass 33 people continuously in an average minute, that is, one person in about 2 seconds; According to statistics, if more than 20 people can pass by using subway cards, QR codes, etc., one person can pass in less than 3 seconds

it can be seen from the data comparison that the speed has not been significantly improved, and the accuracy of face recognition will fluctuate and be unstable under the influence of variables such as light and skin color. More importantly, in such scenarios and transaction scenarios, face recognition will still be one of many payment options in the long-term future, and will not become the only option. Not to mention the fake need to brush your face to throw garbage and get toilet paper

why does the United States show such a refusal when bans are frequent

compared with actively promoting the implementation of face recognition technology in China, the way the United States treats this technology is very extreme. In May this year, the city of San Francisco officially issued a ban on face recognition technology: prohibiting the use of this technology in government agencies and law enforcement agencies. More and more cities such as Oakland, California and Massachusetts are also considering banning or suspending the use of this technology. In early June, Microsoft announced the deletion of its public face recognition database MS celeb

objectively speaking, face recognition technology in the United States has always been at the forefront of industry research, and has long made moves at the commercial level. However, face recognition technology can not achieve 100% accuracy, and the accuracy of different races varies greatly, which means that when different races pass the security check through face recognition technology, some people of color will be more tested. This is a politically incorrect thing in the United States

in addition to questioning the technology itself, Americans are more concerned about the hidden dangers of privacy security. The support behind face recognition technology is massive data. If private data is stolen, your face will no longer be your face, and may belong to anyone. And with smart speakers gradually entering more families, Amazon echo series and Google home series, which have the highest shipments in the world, frequently violate privacy and security, and there is nothing to hide from voices and portraits, which makes Americans anxious

the refusal attitude of Americans also shows a new situation: while changing from weak AI to strong AI, more and more people have shifted from focusing on the development of technology itself to considering more factors related to humanities and ethics. Where is the boundary of artificial intelligence application, and how to improve the transparency of the algorithm has also become a problem that must be solved in front of enterprises and countries

first of all, although China's network security law clearly includes personal biometric information into the scope of personal information, the use, storage, transportation and management of information still need to be further refined

secondly, industry norms should be developed simultaneously with legal norms. For enterprises in subdivided fields, it is necessary to further clarify standards, strictly enforce market access, and avoid industrial chaos from the source

thirdly, define the application boundary of face recognition technology and give full play to the maximum efficiency of the technology itself. Face recognition technology is the most widely used in the field of security. In 2018, the total output value of the security industry reached 718.3 billion yuan, accounting for 0.8% of the total domestic GDP. It occupies a decisive position in terms of market size and social importance. Face recognition technology used in arresting fugitives, screening illegal persons, investigating missing persons and many other aspects will greatly save the cost of public security personnel, more convenient and efficient to fight crime and serve the masses. Standing at the current time node, this application becomes very important

how can computer vision companies seize the market under the pattern of "four super and many strong n-following"

for decades, computer vision, as the technology with the deepest basic research in artificial intelligence system, has always been a topic of interest and research. It is also one of the most widely used technologies in the quality inspection of building materials, such as Jinan gold assay universal material testing machine, pressure testing machine, bending resistance testing machine, etc. Since 2015, based on China's AI policies and the deep cultivation of AI enterprises, China's face recognition technology has ushered in a period of rapid development, and a large number of excellent start-ups have begun to emerge

it can be seen from the above figure that domestic companies specializing in computer vision have obviously formed a pattern of "four super and many strong n-following" in the development process of recent years

the four super brands are Shangtang technology, Kuangshi technology, Yitu technology and Yuncong technology, which are also recognized as the CV four dragons in China. From the chart, we can see that these four companies have gone deep into the security market and formed fierce market competition. According to the 2018 China security industry survey report released by CPS Zhongan United Shangtang technology, the overall market share of computer vision enterprises represented by these four companies in the security field has reached 69.4%

however, the current security market is still facing competition from Hikvision, Dahua and other enterprises. At the same time, there are still computing bottlenecks and a large number of unstructured and semi-structured data preprocessing problems in the industry, which are better requirements for the CV four dragons

how strong refers to star start-ups such as Geling deep pupil and panlite polycarbonate, the fourth emperor, whose impact resistance is 200 times that of glass. Their development strength should not be underestimated. Take the fourth paradigm as an example. As the only start-up enterprise jointly invested by China's five major banks, more than half of the important domestic state-owned banks and national joint-stock banks are customers of the fourth paradigm. At present, the asset scale of financial institutions they serve is more than 50 trillion, and the share of head financial customers is more than 70%. Catching up with and surpassing the international level in technology and relatively accurate market positioning have become the typical characteristics of this group of start-ups

n follow refers to a large number of AI start-ups like Zhongke Huiyuan and videolanguage technology, which have sprung up in the past two years. Their application scenarios are more detailed, and they also begin to consciously avoid the highly competitive security, financial and other markets. Zhongkehuiyuan cut into the field of cover glass inspection in industrial appearance inspection, while Shiyu technology chose the field of visual inspection in the textile industry to solve the pain points of the industry in different ways, so as to reduce costs and increase efficiency for enterprises

in the landing process of computer vision technology, it can be found that algorithms and data are no longer barriers to current competition. Only by finding application scenarios deeply coupled with technology in the vertical field and deeply mining, can enterprises really go a long way. On this basis, truly benefiting people is the direction that all science and technology enterprises should consider

is the ubiquitous face recognition just needed or a gimmick? It is collated and released by China mechanical and electrical products trading (hereinafter referred to as machine trading). If you need to reprint, please indicate the source of the article. For more information about the electromechanical industry, please click attention: electromechanical information of China's electromechanical products trading

Copyright © 2011 JIN SHI