With the rapid development of the security field, face recognition access control machine is also widely used in the security field, has long replaced the traditional security system, do not need to bring keys, access cards, as long as the brush a face on the line, convenient, fast and trouble-free, face recognition door for high-end buildings, so how much do you know about face recognition access control system?
1. Live detection
Live detection function, is to determine whether the current recognition area in the current is a living creature, because some people may use photos to fool the machine, and real-time testing can effectively resist common attacks, such as photos, facial changes, masks, masking and screen remapping.
There are two types of live detection, one is matching type; the second is non-matching type. Matching type is where people need to perform a specified action on demand, such as blinking, while non-matching type does not need to perform any action.
Generally speaking, the non-matching type can be used without coordination; after all, it is both convenient and labor-saving, but the non-matching type has certain requirements on the hardware and algorithm of the face recognition gate.
2. Scene correspondence
People and scenes are variable, so the face recognition access control must take into account external factors, the flow of people and so on. In order to cope with the complex environment, the face recognition technology used must support a variety of complex environments, such as strong light, low light, night backlight, etc., and can detect various angles of facial position, such as front and side, etc.
Only in this way can the requirements of access control be met and the efficiency of face recognition access control system be improved. If the face recognition door is placed outdoors, the requirements for this function will be very high when there is strong interference from sunlight, thunder and rain.
3. Algorithm deployment
Generally speaking, the face recognition algorithm in the face recognition access control will be deployed on the cloud server or locally, if deployed in the cloud, the hardware configuration requirement for the is not high, in order to reduce the hardware cost, the face recognition algorithm can be deployed in the server, the result of this is not because of power failure, network disconnection and other emergency, face recognition is impossible, the data may be lost, the best choice is in the machine The face recognition algorithm is deployed locally on the machine, even if it is offline, it will not affect the use and also protect the local data to avoid data loss.