Facial recognition payment is coming, QR code payment comes to an end?

进入2019年，我国支付行业发展相对平静。是静水流深，还是爆发前的宁静？但我们看到，刷脸支付已经暗流汹涌，从布局到发力，是否会引起新一轮的各方角力和市场变局呢？

Entering 2019, Chinese payment industry has developed relatively quietly. Is it the still water flowing deep, or is it quiet before the explosion? However, we have seen that facial recognition payment has been undercurrent. Will it cause a new round of competition and market changes from layout to breaking out.

一、刷脸支付的发展行进路线

The development route of facial recognition payment

从技术储备来说，生物识别技术有一个逐步成熟的过程。特别是人脸识别技术在最开始具有不小的争议。随着3D结构光学技术、大数据和AI技术的进步，人脸识别技术精准度持续提升，进而能够达到商用的标准。虽然并没有一个人脸识别技术公认的差错率门槛，但是大家对人脸识别技术的信心越来越足，刷脸技术应用场景不断扩张和深入，其在身份验证链条上大有从辅助地位“上位”到主体位置的态势。

From the perspective of technical reserves, biometric has a process of maturity. In particular, face recognition technology has a lot of controversy at the beginning. With the advancement of 3D structured light camera, big data and AI technology, the accuracy of face recognition technology continue to increase, so it can reach commercial standards. Although there is no recognized error rate threshold for face recognition technology, everyone's confidence in face recognition technology is getting firmer. The application of face recognition technology is expanding and developing, and it has a large tending from auxiliary to leading position in the authentication chain.

支付宝是行业内最早布局人脸识别技术的公司之一。自2015年起，支付宝率先将人脸识别技术应用于用户登录后，这一技术先后用于实名认证、找回密码、支付风险校验等场景。2018年8月15日，支付宝宣布在经过经验积累和技术升级之后，刷脸支付已经具备了商业化的能力，在未来一年内将向各种商业场景普及自助收银+刷脸支付的解决方案。2018年9月1日，支付宝在肯德基的KPRO餐厅上线刷脸支付，不用手机，通过刷脸即可支付，这也是刷脸支付在全球范围内的首次商用试点。从2015年到2018年，花了三年时间。自此，刷脸支付进入了快速发展时期。2018年12月13日，支付宝发布了一款叫做“蜻蜓”的刷脸支付产品，直接将刷脸支付的接入成本降低了80%，这意味着刷脸支付真正具备了大规模普及应用的前提。2019年4月17日，支付宝宣布推出第二代基于线下消费场景的刷脸支付机具“蜻蜓”，定价1999元，相比第一代直降近30%，这也标志着线下刷脸支付进入“千元机”时代。2019年5月9日马化腾本人也在北京万达广场体验微信刷脸支付，亲自为自家刷脸设备站台。一些支付机构也开始试点线下刷脸支付设备。同时，在百度输入刷脸支付，热度逐渐升温，而且也有很多线下代理商开始进入这个市场。

Alipay is one of the first companies in the industry to deploy face recognition technology. Since 2015, Alipay has taken the lead in applying face recognition technology to users login. This technology has been used for real-name authentication, password recovery, and payment risk verification. On August 15th, 2018, Alipay announced that after the accumulation of experience and technology upgrades, the face payment has already been commercialized, and in the next year, it will popularize the self-service cash register and face payment solution for various business scenarios. On September 1st, 2018, Alipay launched facial recognition payment at the KFC’s KPRO restaurant. You can pay by scanning your face without using a mobile phone. This is the first commercial trial place of face payment in worldwide. It took three years from 2015 to 2018. Since then, face payment has entered a period of rapid development. On December 13, 2018, Alipay released a face-pay product called “Qingting”, which directly reduced the access cost of face payment by 80%, which means that face payment really has a large-scale popular application premise. On April 17, 2019, Alipay announced the launch of the second generation of face payment device "Qingting", based on the offline consumption scene. The price is 1999 RMB, which is nearly 30% lower than the first generation. This also marks the offline face payment device into the "thousand yuan machine" era. On May 9th, 2019, Pony Ma himself also experienced WeChat face payment at Beijing Wanda Plaza, and personally scan his own face for his own device. Some payment agencies have also begun to try to R&D offline payment devices. At the same time, the heat of face payment in Baidu search engine gradually warmed up, and there are many offline agents began to enter this market.

中国银联在2018年12月于北京和上海小规模推行了刷脸支付服务。持卡人只需在云闪付APP上注册并开通“刷脸支付”服务，选择一张银联卡作为默认支付卡，即可在指定门店抢先体验全新的支付服务。2019年4月底，银联再次对外宣布刷脸支付项目进度，在天津现代职业技术学院试点刷脸支付。2018年12月，宁波率先启动央行标准线下刷脸支付试点验证工作。在不突破现有业务规则的前提下，利用机器学习、图像识别、密码技术等科技手段提升用户支付体验，增强交易安全防护能力。

China UnionPay launched a small face payment service in Beijing and Shanghai in December 2018. Cardholders only need to register on the UnionPay app and open the “Face Payment” service, select a UnionPay card as the default payment card to experience the new payment service in the designated store. At the end of April 2019, UnionPay once again announced the progress of the face payment project, and start face payment trial at the Tianjin Modern Vocational and Technical College. In December 2018, Ningbo city took the lead in launching the trial verification of the face payment under the central bank's standard line. Based on the premise of not breaking the existing business rules, using machine learning, image recognition, cryptography and other scientific technological means to enhance the user's payment experience, enhance transaction security protection capabilities.

银行方面，招商银行是国内首家研究人脸识别技术的银行。目前，已相继推出了刷脸取款、刷脸转账、刷脸支付等功能。光大银行的人脸识别技术也应用于账户登陆、转账、线上融资、线上申请办卡等不同场景。农行正全国推行刷脸取款，建行正商用刷脸支付。值得注意的是，多数银行手机APP为了确保交易的安全可靠，登陆手机APP可以单独应用人脸识别进行验证，但若用户运用银行手机APP进行转账等交易，并不会单独应用人脸识别，而是与密码等其他验证方式共同进行认证。

In terms of banks, China Merchants Bank is the first bank in China to study face recognition technology. At present, the functions of face recognition withdrawal, face recognition transfer and face payment have been introduced by CMB. China Everbright Bank’s face recognition technology is also used in different scenarios such as account login, transfer, online financing, and online application for card. Agricultural Bank of China is pursuing a face-lifting fundraising across the country, and CCB is preparing to using face payments. It is worth noting that in order to ensure the security and reliability of transactions, most mobile phone bank APP can use face recognition for verification individually. However, if a user uses a bank mobile phone APP to conduct transactions such as transfer, face recognition will not be applied separately. It is certified together with other authentication methods such as passwords.

二、刷脸支付的体验比对

Face payment experience comparison

一般来说，刷脸支付在近场支付的商用价值要远远高于远程支付。在线下场景中，我们把刷脸支付与二维码支付做比对。

In general, the commercial value of face payments in near-field payments is much higher than remote payments. In the online scene, we compare the face payment with the QR code payment.

（一）刷脸支付在效率上具备一定优势。在自助收银机上从结账到刷脸支付整个过程耗时不超过10秒，也免去了排队等待等环节，大大节约了时间。在与二维码支付的比对上，二维码支付需要调出二维码，增加了支付环节。但是在实际场景中，由于消费者可以提前调出二维码，因此效率差距可能并不大。

Face payment has certain advantages in efficiency. In the self-service cash register, the entire process from checkout to making payment takes no more than 10 seconds, and the queue waiting is eliminated, which greatly saves time. In the comparison with the QR code payment, the QR code payment needs to generate the QR code, which increases the payment process. However, in the actual scenario, since the consumer can generate the QR code in advance, the efficiency gap may not be large.

（二）刷脸支付体验上具有一定独特性。在体验上，从笔者感受而言，单就刷脸支付的体验似乎更有高级感和科技因素在里面，虽然这种感受因人而异，但可能对年轻人更具吸引力。就基于手机的二维码支付而言，刷脸支付解放了双手，不需要腾出手来拿手机支付；如果仅凭刷脸而言，如果前期做好人脸认证、双免以及限额设置，可能刷脸支付更为便捷，对于老年人等群体更为友好。

The face payment has certain unique experience. In terms of the author's feelings, the experience of face payment is more high-level and more technological. Although this feeling varies from person to person, it may be more attractive to young people. In terms of mobile phone-based QR code payment, the face payment liberates the hands, and does not need to free up the hand to operate the mobile phone to pay; For the face payment, if the face authentication, double-free and limit settings are done in advance, it is more convenient and more friendly to groups such as the elderly.

（三）刷脸支付在使用上具有一定独立性。由于不需要手机，那么就不用担心手机没电、断网，可能这是很多消费者手机支付的痛点。我们可以看到，用手机进地铁，出地铁时手机没电刷不了二维码或者NFC，在地铁里时有发生。同时，5月底，上海地区移动网络出现异常，移动数据网络无法连接，必然也会对移动支付造成影响。刷脸支付摆脱了手机制约，不需要手机就能支付，支付自由度大大提升。

Face payment has a certain degree of independence in use. Since there is no need for a mobile phone, there is no need to worry about the lack of power or network disconnection of the mobile phone. This may be a pain point for many mobile phones pay consumer. We can see that when using the mobile phone to enter the subway, and the mobile phone is power off and can't scan the QR code or NFC to get out the subway, this happens sometime in the subway. At the end of May 2019, the mobile network in Shanghai was abnormal, and the mobile data network could not be connected, which inevitably affect mobile payment. The face payment is free from the restrictions of the mobile phone, and can be paid without a mobile phone, and the freedom of payment is greatly improved.

三、笼罩刷脸支付的几个风险问题

Several risk issues for face payment.

（一）刷脸支付的人脸识别技术问题。相较于以前，识别技术一直在进步。特别是人脸识别算法逐步成熟，其1：1比对的误识率已可低于十万分之一，足以达到商用的标准。关于长相相近的人的误识问题，包括刷脸中的“双胞胎”问题，还不能很好解决。人脸识别技术的应用难点在于支付环节的应用安全性要求更高、线下场景更为复杂以及公开环境、公共设备的挑战更大。

Face recognition technology issues for face payment. Identification technology has been improving compared to the past. In particular, the face recognition algorithm has gradually matured, and the mis-recognition rate of the 1:1 comparison can be less than one in 100,000, which is enough to meet the commercial standard. The mis-recognition of people with similar looks, including the "twins" face payment is not well resolved. The application difficulty of face recognition technology is that the application security requirements of the payment link are higher, the offline scene is more complicated, and the open environment and public equipment are more challenging.

需要指出的是，刷脸支付的安全性并不会有质的提升。因为第一次的人脸认证，依然依赖于原有的认证体系，如公安部身份证信息系统、银行账户认证等。因此，办身份证或者开立银行账户时的现场核验，依然是身份核实的基础。我们可以看到，很多高铁站和机场把人脸识别与身份证的人脸做比对，来确认身份证与人之间的一致性。人脸认证也是基于此逻辑，人脸的首次认证也是人脸信息收集的行为，是刷脸支付的一个前置条件。每一次刷脸支付是人脸识别与第一次人脸认证的信息做比对，只有客户和账户形成唯一性的对应后，才形成后续的支付行为。

It should be pointed out that the security of face payment does not have a qualitative improvement. Because the first face authentication, still based on the original certification system, such as the Ministry of Public Security ID card information system, bank account certification. Therefore, on-site verification when registering an ID card or opening a bank account is still the basis for identity verification. We can see that many high-speed rail stations and airports compare face recognition with the face of an ID card to confirm the identity between the ID card and the person. Face authentication is also based on this logic. The first authentication of the face is also the behavior of face information collection, which is a precondition for face payment. Each face payment is compared with the information of the first face authentication. Only when the client and the account form a unique correspondence, the subsequent payment behavior is formed.

（二）刷脸支付的“隔空盗刷”问题。银行卡需要刷，二维码需要扫，而人脸直接识别。看似越来越简单。但是我们可以看到，银行卡只有通过侧录或者录像方式盗取卡号和密码。二维码则出现收款码被隔空盗刷，被人扫码的情况。这种情况虽然很少见，但是目前还没有很好的手段去直接防范。

Money stolen in face payment. The bank card needs to be swiped, the QR code needs to be scanned, and the face is directly recognized. It seems to be getting simpler and easier. However, we can see that the card number and password of bank card might be stolen only through side recording or video recording. The receipt code is stolen and scanned by other people on site while using QR code pay. Although this situation is rare, there is currently no good means to prevent it directly.

由于刷脸支付屏蔽了银行卡、二维码等支付工具或者介质，人脸识别直接成为关联账户和解锁密码的凭证。在免密免签的情况下，只要通过刷脸进行身份识别后，就可以直接扣划资金。而刷脸支付的“隔空盗刷”，似乎会更方便。因为二维码的“隔空盗刷”有时机的限制，需要被害人从手机中调出二维码来。而刷脸支付则不存在这个问题，在双免的情况下，如果技术能够解决远距离识别的问题，那么人脸将成为不设防的“收款码”，犯罪分子可以轻易盗刷。在安全和便捷之间成为艰难的抉择。如果不能解决这个问题，将成为笼罩在刷脸支付上的乌云。

Since the face payment blocks the medium such as a bank card or a QR code, the face recognition directly becomes a certificate for the associated account and the unlock password. In the case of a password-free and signature-free, the funds can be directly deducted after identified by face recognition. It seems more convenient for criminal to steal money by face payment. Money stolen by QR code pay is limited by the time limit, it needs the victim to retrieve the QR code from the mobile phone. However, there is no such problem in face payment. In the case of double-exemption, if the technology can solve the problem of remote recognition, the face will become an unprotected "receipt code", and criminals can easily steal it. It is a tough choice between safety and convenience. If you can't solve this problem, it will become a blemish for the face payment.

针对刷脸盗刷钱的问题有两个解决方案：一是要想解决可能出现的“隔空盗刷”问题，那么就要压缩刷脸支付扫描的距离，现在可以做到0.3-3米都能实现3D扫描，如果能把距离定在0.3米以内，形成技术强制性规范可以防范这个问题。问题在于考虑到当前对扫描终端管控力度，如果技术上实现3米的3D扫描，那么，犯罪分子铤而走险的可能性会很大，而且只要解决技术上的问题，“隔空盗刷”实施起来可能比二维码更容易。当然，这只是笔者的一种假设。另外一个解决方案是刷脸配套输入密码，但是这个体验就会很差。

There are two solutions to the problem of money stolen by face payment: First, To solve the problem of money stolen, it needs to compress the distance of face recognition. Now it can realize face recognition in 0.3-3.0 meters, if the distance can be set within 0.3 meters, and this formation of technical mandatory specifications can prevent “money stolen” problem. The problem is that considering the current control of the scanning terminal, if the technical realize 3D scanning in 3 meters, the possibility of criminals taking risks would be higher, and as long as the technical problems are solved, money stolen may be more implemented. It's easier than QR code pay. Of course, this is just an assumption of the author. Another solution is to enter the password while use face payment but the experience will be poor.

（三）关于刷脸支付个人隐私信息保护的问题。

Private secret info protect issues in face payment

一是人脸生物信息的泄露问题。由于人脸信息，其固有特性、采集方式、集中存储的特点导致信息泄露风险较大。生物特征数据存储集中度越来越高。一旦热点应用的生物特征库被攻破，极易导致大规模隐私泄露，甚至会引发系统性风险。二是由于脸部识别信息具有唯一性和不可更改性。相比于密码泄露或疑似泄露后，可以随时更改。但是传说的“换脸”或者“变脸”不存在。三是任何一个数据进入到计算机里面以后，都会变成计算机代码，生物特征数据也不例外。如果由于个人生物信息泄露，犯罪分子掌握了被害人的脸部识别信息，一旦这些数据被还原，并被黑客等犯罪分子拿到后，是否可以直接获取被害人的账户和密码，从而轻易地盗取资金。

The first is the leakage of biological face information. The face information has its inherent characteristics, collection methods, and centralized storage characteristics, which lead to a greater risk of information leakage. Biometric data storage is becoming more and more concentrated. Once the hot-applying biometric library is compromised, it can easily lead to large-scale privacy breaches and even systemic risks. The second is because the face recognition information is unique and unchangeable. It’s not like the traditional security method, which can be changed at any time after a password leak or suspected leak. The legendary "face change" does not exist. Third, any data that enters the computer will become a computer code, and biometric data is no exception. If the personal biometric information is leaked, the criminals have mastered the victim's face identification information. Once the data is restored and obtained by criminals such as hackers, can the employee's account and password be directly obtained, so that the funds can be easily stolen.

从现在情况看，刷脸支付最大的问题在于人们对于个人隐私信息的担忧，成为刷脸支付发展进程中的“灰犀牛”。用户生物特征普遍暴露在商场、饭店等各种公共场所，不法分子可通过远程、非接触方式，在用户本人毫无察觉的情况下“无声无息”地非法批量采集生物特征信息。刷脸支付的普遍应用，可能导致个人生物的非法采集和买卖成为新的“黑产业链”。这将引发人们对于个人隐私被非法收集和储存的担忧。这种担忧将比之前的个人隐私更甚。微软目前已经悄然删除其最大的公开人脸识别数据库——MS Celeb。MS Celeb数据库2016年建立，拥有超过1000万张图像，将近10万人的面部信息。微软描述其为世界上最大的公开面部识别数据集，并用于培训全球科技公司和军事研究人员的面部识别系统。

From the current situation, the biggest problem with face payment is that people's concerns about personal privacy information have become the "Grey rhinoceros" in the development process of face payment. User’s biometric is generally exposed to various public places such as shopping malls and restaurants. The criminals can illegally collect biometric information in a remote and non-contact manner without any notice from the user. The widespread application of face payment may cause the illegal collection and sale of personal creatures as a new “black industry chain”. This will raise people’s concerns about the illegal collection and storage of personal privacy. This concern will be even worse than the previous personal privacy. Microsoft has quietly removed its largest public face recognition database, MS Celeb. The MS Celeb database was established in 2016 and has over 10 million images and nearly 100,000 facial information. Microsoft describes it as the world's largest public facial recognition data and is used to train facial recognition systems for global technology companies and military researchers.

人脸识别技术还引发过其他争议。例如，它使得跟踪个人行踪和探查个人行为变得轻而易举，从而对个人隐私造成大规模的侵害。如今一些执法机构经常依靠这项技术来帮助调查案件。如何在保证城市生活安全和个人隐私之间做到平衡，成为许多城市管理者需要解决的问题。今年5月份，美国旧金山就出台了一项法令，禁止警察和其他政府机构使用人脸识别技术。

可以预见，关于人脸识别技术应用产生的个人隐私问题引发的争议将会延绵不绝，可能对刷脸支付的发展前景蒙上一层阴影。

Face recognition technology has also caused other controversies. For example, it makes it easy to track personal whereabouts and detect personal behavior, thereby causing massive violations of personal privacy. Today, some law enforcement agencies often rely on this technology to help investigate cases. How to balance the security of urban life and personal privacy has become a problem that many city managers need to solve. In May of this year, San Francisco in the United States issued a decree prohibiting police and other government agencies from using face recognition technology.

It is foreseeable that the controversy over the personal privacy issues arising from the application of face recognition technology will continue to linger, which may cast a shadow over the development prospects of face payment.

四、刷脸支付的市场“赌局”

Gambling in face payment market.

目前，对于将人脸识别技术应用于支付业务，政策上有一定松动。2015年末，央行公布了《中国人民银行关于改进个人银行账户服务加强账户管理的通知》，提出“提供个人银行账户开立服务时，有条件的银行可探索将生物特征识别技术和其他安全有效的技术手段作为核验开户申请人身份信息的辅助手段”。2017年，中国人民银行发布《中国人民银行关于优化企业开户服务的指导意见》，鼓励银行积极运用技术手段提升账户审核水平，包括鼓励银行将人脸识别、光学字符识别(OCR)、二维码等技术手段嵌入开户业务流程，作为读取、收集以及核验客户身份信息和开户业务处理的辅助手段。但这里都强调的是生物特征识别技术作为开户业务处理的辅助手段。如果刷脸技术来替代身份识别或者密码的验证，甚至于远程开户，在政策上似乎还没有“开闸”。

At present, there is a certain loosening of the policy for applying face recognition technology to payment services. At the end of 2015, the central bank announced the "Notice of the People's Bank of China on Improving Personal Bank Account Services to Strengthen Account Management", proposing that "when providing personal bank account opening services, conditional banks can explore biometric and other safe and effective technical means as an auxiliary means to verify the identity information of the applicant. In 2017, the People's Bank of China issued the “Guiding Opinions of the People's Bank of China on Optimizing Enterprise Account Opening Services” to encourage banks to actively use technical means to improve account review, including encouraging banks to identify faces, optical character recognition (OCR), and QR codes. The technical means are embedded in the account opening business process as an auxiliary means for reading, collecting and verifying customer identity information and account opening business processing. However, the emphasis here is taking biometric as an aid to the processing of account opening. If the face recognition technology is used instead of the identification or password verification, even open account remotely, it seems that there is no “opening” in the policy.

在监管部门态度未明的情况下，支付宝有着“all in”的决心和勇气。因为在二维码支付的赛道上，支付宝不具有优势。而微信支付的杀器就是依托微信这一强大的社交软件。因此，支付宝必须换一条赛道。刷脸支付由于不依赖于手机，摆脱了微信强大的优势，有助于其在零售支付领域重新获得新的优势。

With the unclear attitude of the regulator, Alipay has the “all in” determination and courage. Because on the track of QR code pay, Alipay does not have the advantage. WeChat payment is based on WeChat, a powerful social software. Therefore, Alipay must change a track. Face payment is not based on mobile phones, gets rid of the strong advantages of WeChat and helps it regain new advantages in the field of retail payment.

因此，支付宝在刷脸支付开始投入巨额资金进行推广。据“2019中国智慧商业信息化展览会”上的展商介绍，支付宝的推广策略是：将“蜻蜓”原价卖给商家，商家激活设备产生第一笔有效交易后即可获得240元补贴。此后连续三十天，按照每日设备上产生的有效去重用户数进行补贴奖励，每日激励金额上限为240元。也就是说，商家花2688元购买这款设备后，最多可获得7200元补贴，最后净赚近5000元。

Therefore, Alipay began to invest huge sums of money in the face payment. According to the exhibitors at the "2019 China Smart Business Informatization Exhibition", Alipay’s promotion strategy is: sell “Qingting” to the merchant at a original price, the merchant activates the equipment to generate the first effective transaction and then get a subsidy of 240 yuan. After 30 consecutive days, the subsidy will be awarded according to the number of valid de-duplicated users generated on the daily equipment. The maximum daily incentive amount is 240 yuan. In other words, after the merchant spends 2,688 yuan to purchase this equipment, it can get a subsidy of up to 7,200 yuan, and the net profit is nearly 5,000 yuan.

微信支付反而显得有些迟疑。当然，既然对家出手，肯定要跟，但是跟多少就是个选择了。腾讯祭出了其一贯的跟随策略。可能腾讯觉得现在并不是大规模跟进的时机。毕竟刷脸支付的政策前景还是具有一定不确定性，培养刷脸支付习惯的投入产出比还不确定，或者腾讯相信二维码支付的优势短期内还无法被动摇。但是，这个跟随策略在刷脸支付这条赛道上是否能奏效呢？当前，相较于二维码支付，刷脸支付的硬件成本还是比较高，而且占空间。因此，其应用场景没有二维码支付丰富。我们可以观察，在商超、餐饮以及便利店等场景中，收银空间是有限的，布放哪怕像PAD大小的刷脸支付设备都会觉得局促。如果同时布放两个哪怕像PAD大小的设备都会有困难。所以，在这种情况下，似乎先发优势变得更为重要。

WeChat payment seems to be somewhat hesitant. Of course, since your competitor has took a bet, you must follow it, but how much is a choice. Tencent takes its consistent follow-up strategy. Tencent may think that it is not the time to follow up on a large scale. After all, the policy prospects for face payment are still somewhat uncertain. The input-output rate of developing face payment habits is still uncertain, or Tencent believes that the advantages of QR code payment cannot be shaken in the short term. However, does this follow-up strategy work on the face payment track? Currently, compared to QR code payment, the hardware cost of face payment is still relatively high, and it takes up space. Therefore, its application scenario is not rich as QR code payment. We can observe that in the scenes of supermarkets, restaurants and convenience stores, the cashier space is limited, and even the face-paying devices like PAD will feel cramped. It would be difficult to deploy two devices that are even the size of a PAD at the same time. So, in this case, it seems that the first mover advantage becomes more important.

二维码支付时代，几乎每家门店都同时张贴着微信和支付宝的收钱码，或者干脆是聚合支付二维码。这个画面在刷脸支付时代也许将不复存在。有观点认为，刷脸支付是排他性的，很难做聚合支付。二维码时代的选择权在消费者手中，取决于打个哪个应用，而刷脸支付的选择权掌握在商家手里。

In the era of QR code payment, almost every store has posted the code of WeChat and Alipay at the same time, or simply aggregated payment QR code. This picture may no longer exist in the era of face payment. There is a view that face payment is exclusive and it is difficult to make aggregate payments. The choice of the QR code era is in the hands of consumers, about which payment method to choose, and the choice of face payment is in the hands of the merchant.

刷脸支付会闭环特征更加明显，联通的难度在变大。银行卡是联网通用，二维码支付可以联码通用，到了刷脸支付，由于支付介质直接变成生物特征，不再是介质的聚合，而可能成为识别终端聚合，一个硬件聚合不同的识别体系和信息渠道，然后进行路由选择不同的渠道。但是硬件的聚合可能更加困难。线下刷脸的前提是用户在线上的专门软件中完成免密支付开通以及人脸信息录入。在这一大前提之下，硬件设备的使用可以成为软件下载的导流，但软件的普及率也相应会成为商家选择硬件时的重要考量。

The closed-end feature of the face payment is more obvious,and the difficulty of uniting is getting bigger. The bank card is universal after connecting to the network, and the QR code payment can be commonly used in the universal code. The payment medium directly becomes a biometric in face payment, it is no longer the aggregation of the medium, but may become the identification terminal aggregation, and a hardware aggregation different recognition system and information channels, then routing different channels. But the aggregation of hardware can be more difficult. The premise of offline face payment is to complete the password-free payment and face information verification in the special software by user online. Under this premise, the use of hardware devices can become a guide for software downloads, but the popularity of software will also become an important consideration when choosing hardware.

大概率的事件是重新引发新一轮的圈地大战，大家都想锁定商户，最后当市场格局稳定下来之后，无论是监管引导还是市场自发，会形成支付聚合，防止出现“一柜多机”和消费者被“绑架”等问题。

The high probability event is to reignite a new round of enclosure wars, everyone wants to lock more merchants. When the market structure stabilizes, whether it is regulatory guidance or market spontaneous, it will form a payment aggregation to prevent the emergence of “one cabinet and multiple machines” and consumers are being "kidnapped" and other issues.

还需要考虑的另外一个问题在于，刷脸支付的习惯培养是否可以向二维码支付那样用钱砸出来呢？我们认为是可以的。但是，可以不等于不计代价。可能培养刷脸支付的习惯花费的代价要高于扫码支付。这就会涉及到市场主体之间的博弈。

Another question that needs to be considered is whether the habit of face payment can be raised by money like QR code payment. We think it can work. However, workable doesn’t mean at any cost. It may cost more money to cultivate the habit of face payment than QR code payment. This will involve a game between market players.

五、相关建议

（一）加快建立人脸识别技术的标准和业务规范。如3D结构光的扫描距离、刷脸支付的支付限额以及刷脸支付应用的禁入环节或场景等等。

Accelerate the establishment of standards and business specifications for face recognition technology. Such as the scanning distance of 3D structured light camera, the payment limit of face payment, and the forbidden link or scene of face payment application.

（二）建立对刷卡支付受理终端的检测和认证。由于受理终端实现人脸识别，涉及到生物信息收集，需要加强对隐私信息的保护和控制。这就需要对受理终端的软硬件标准加强检测和认证，明确其入网和使用门槛。

Establish detection and certification for the card payment acceptance terminal. Since the acceptance terminal implements face recognition and involves the collection of biological information, it is necessary to strengthen the protection and control of the private information. This requires enhanced testing and certification of the hardware and software standards of the acceptance terminal, and clarifies the threshold for access and use.

（三）对于市场机构收集人脸生物信息要加强管控。特别是人脸信息的传输和储存要提高安全等级。建议对生物特征信息进行加密传输和加密储存。

Strengthening control over the collection of facial biological information by market institutions. In particular, the transmission and storage security level of face information should be enhanced . It is recommended to encrypt the transmission and storage of biometric information.

（四）建议推动建立国家级的生物特征信息管理平台。要评估某个市场机构掌握几亿用户人脸信息对用户以及国家信息安全的影响，推动基于分布式和加密技术的生物特征信息管理平台，为生物识别提供统一的验证服务。

It is recommended to promote the establishment of a national-level biometric information management platform. To evaluate the impact of hundreds of millions of user face information for users and national information security, a bio-information information management platform based on distributed and encryption technology is providing a unified verification service for biometric identification.

（五）推动刷脸支付的开放和聚合。刷脸技术的核心是算法，要推动刷脸支付底层技术的开放和算法的输出。同时在刷脸支付的受理端，要推动硬件聚合和软件互通，要防止出现垄断和闭环。

Promote the openness and aggregation of face payment. The core of face recognition technology is the algorithm, which is to promote the openness of the face recognition underlying technology and the output of the algorithm. Meanwhile in the acceptance face payment terminal, it is necessary to promote hardware aggregation and software interoperability, and to prevent monopoly and closed loop.

总之，刷脸支付很有意思，具有很高的话题性和市场颠覆能力，比如刷脸支付如果普及，是否会成为公安部天眼系统的“民兵”，犯罪分子可能被淹没在人战争的汪洋大海中。因为可能到时有几千万个人脸识别终端，你买个东西都要人脸识别，犯罪分子可能哭晕在角落里。但是刷脸支付，涉及到人脸识别，存在的个人隐私信息保护的问题将挥之不去。因此，刷脸支付的政策前景依然不太明朗，但在几家巨头的竞争和推动下，市场前景看好。刷脸支付未来将在争议中前进，在规范中完善，在创新中发展。刷脸支付能否刷出零售支付新热度，让我们拭目以待。

In short, face payment is very interesting, with high topicality and market subversive ability. For example, if the face payment is popular, will it become a “militia” of the Ministry of Public Security’s Sky System, and criminals may be submerged in the sea of people’s war. Because there may be tens of millions of face recognition terminals, you need to use face recognition when you buy something, and criminals may cry in the corner. However, the face payment involve face recognition, the problem of the protection of personal privacy information will be lingering. Therefore, the policy for face payment is still unclear, but with the competition and promotion of several giants, the market prospects are promising. The face payment will advance in the controversy, improved in the norms, and developed in innovation. Whether face payment can bring the new hot-spot in retail payment, let us wait and see.