Since alphago defeated Li Shishi in 2016, AI has completely recovered from hibernation and rapidly ushered in new opportunities for industrial development. Due to the characteristics of big data and large-scale, the security field has become the best application market for AI. With the support of artificial intelligence, Internet of things, cloud computing, big data and other emerging technologies, the development of security industry has changed from the concept of safe city construction to the concept of smart city, and the boundary of big security has gradually become blurred, which has been extended to the pan security field with video Internet of things as the core.
In the system of artificial intelligence technology, face recognition is undoubtedly a widely used biometric technology. Face recognition technology is an advanced technology that extracts effective feature information from face image and uses computer to analyze it, so as to carry out identity authentication. At present, the research of face recognition technology has become a hot spot in the field of artificial intelligence and recognition. Their products are mainly used in public security and financial fields.
In the past two years, with the rapid iteration of face recognition technology (it is said that there will be a qualitative leap every six months), there are more and more subdivision scenes for face recognition, and gradually from the big security to the civil and commercial application market. It can be said that at this stage, face recognition technology has been everywhere in our lives, such as turnstile gate for apartment, community face brushing, face brushing payment, face brushing attendance and face brushing withdrawal, etc. Because of its non-contact, convenient and reliable, people are gradually accepting this application, although face recognition has certain security risks.
At present, the access control and entrance control system of commercial application scenarios is gradually changing from card authentication to face recognition. “In the past two years, biometric technologies such as face recognition have been gradually popularized in the pan security market. In the application fields of commercial scenes such as apartments, office buildings, hotels and shopping malls, face recognition technology is more popular. The reason is that face recognition is non-contact recognition, and the most important thing is that it can realize multi-target recognition at the same time, greatly improving business efficiency.”
The application of face recognition technology has strong scene, and different application scenarios have different requirements for the core algorithm ability. In the office area, community, apartment, hotel and other pan security business scenes, the face database is relatively small, and it does not need deep learning training on a large number of data sets, and there is no high requirement for hardware, and the embedded face terminal can bear the computational load of face recognition in these scenes.
As we all know, due to the large floating population, long-term rental apartments, public rental housing and hotels bring many security risks and population management problems. The traditional population control method not only consumes a lot of human and material resources, but also is extremely inefficient, and the results are often unsatisfactory. The application of embedded face terminal will be a new method to solve the problems of access control and security management.
Through face recognition security turnstile gate, it can effectively monitor the flow of people in and out of the apartment. Face is not easy to copy, especially with the continuous progress of human face detection technology, ordinary photos and dolls can’t cheat the access control system, and changing the shape, ordinary makeup and face shape will not affect the recognition results too much. Generally speaking, taking face brushing as the “key” of entrance and exit not only brings convenience to people, but also effectively reduces the management cost of apartment population flow.
However, due to the large number of room doors in apartment and hotel industries, the cost is often too high if high-end face recognition terminals are used. The cat’s eye lock based on front-end detection and capture combined with back-end recognition can reduce the cost of the whole face recognition access control, which is convenient for large-scale replication and application in apartments and hotels. And the passive recognition function is also very important, that is, it needs human touch to turn on the face recognition function, which avoids the embarrassment and management trouble of personnel false recognition. Moreover, the cat’s eye lock type face recognition terminal not only greatly reduces the cost, but also solves the problem of power consumption, with high cost performance, and simple installation and later maintenance.
The face recognition terminal will also have great potential in the smart community. Through the face recognition gate, we can not only timely and effectively find outsiders, visitors and surrounding service personnel, do ID card registration and take photos in time, so as to standardize the community life order, create a smart and safe home, but also protect the owners in an all-round way and provide extremely convenient services. There is no doubt that face recognition access control can strengthen the security system of community entrance and exit access control, and assist the community in all-round intelligent management. It can be said that the cooperation between face recognition and smart community is the best partner, and the future smart community will be face recognition community.
In addition, in the field of office and school, the traditional attendance methods are IC card check-in and fingerprint check-in. In the age of intelligence, with the progress of face recognition technology, face recognition has gradually entered people’s field of vision. Due to the solution of the IC card loss, fingerprint broken skin can not play on the card, and improve the attendance management system, so gradually accepted by the market and widely used. It can be said that the appearance of face recognition attendance is not only the result of the progress of biometric technology, but also the inevitable result of today’s social needs.
Compared with the C-end family field, the access control in the public field tends to use face recognition, and the recognition rate is not so high. Generally, enterprises can accept the recognition rate of 99.7%. But for home users, due to the security and privacy risks, it is difficult to accept that face recognition has a little error. In the entrance and exit of public areas such as schools, communities, office buildings or apartments and hotels, video monitoring and anti-theft alarm devices are generally deployed, which is relatively safe. Moreover, the most important thing is that face recognition can realize multi-target recognition at the same time, avoid the Procrastination of queue recognition, and greatly improve the management and business efficiency.
However, although face recognition has been accepted in the commercial field, high cost and high power consumption have been the obstacles to its large-scale application. Face recognition technology and equipment providers need to meet the needs of practical application scenarios, and at the same time, continuously reduce the cost of equipment, reduce the power consumption of products, and improve the whole life cycle of products, so as to speed up the large-scale application of face recognition technology in the commercial field.
In addition to the cost and power consumption, face recognition also faces the technical bottlenecks of twin recognition, light interference, beauty and makeup, occlusion and low resolution recognition. Both manufacturers and users need to face up to the existing difficulties, and constantly make products and technologies more practical. With the establishment and implementation of national and industrial standards, as well as the continuous iteration of face recognition technology, face recognition will be strengthened in anti-counterfeiting, three-dimensional, environmental adaptability, high performance, high speed, and usher in a new round of development boom.