Face recognition technology has gradually penetrated into every aspect of our lives, what principle does a face recognition machine actually rely on to recognize faces? How reliable will this reliance on machines for face recognition be? Let's explore the recognition principle and reliability of face recognition technology.
The face recognition system requires three steps for each recognition: in the first step, the face is imaged; in the second step, information points are acquired randomly on the face, and every three points are connected into a triangle; in the third step, the datamined triangle is compared with the person's facial data already deposited in the database beforehand, which in turn completes the whole recognition process.
A face recognition door lock consists of a face recognition system, and a door lock gate machine connected to it. The face data is analyzed using artificial intelligence technology to build a face recognition model. When a person passes through the channel, the facial information is extracted and matched with the facial information pre-recorded in the face template database, and then controls the channel gate to open the door after the match is successful.
So the question is, can a face recognition machine tell identical twins, triplets or multiples, who look so similar that people can hardly tell them apart with the naked eye, tell them apart accurately? In today's increasingly sophisticated make-up technology, we can often see that many people are simply different before and after make-up, and can the recognition machine still accurately identify them? Can a 3D printed face with no difference in the position and size ratio of the features still be tested by a face recognition system?
The human face consists of localities such as eyes, nose, mouth, chin, etc. The face recognition system first models these localities and the structural relationships between them, and then compares the extracted face image with the already modeled face template, and then judges the identity information of the face based on the similarity.
Through a series of experimental investigations, it was found that the material of the 3D printed human face is ceramic powder, which cannot effectively reflect the infrared light emitted by the machine, and it appears to be empty in front of the device; the twins who look extremely similar, the face recognition machine can also identify them accurately and quickly; it is also difficult for make-up and disguise to fool the face recognition machine; and those cameras installed in many public places can indeed play the role of identification tracking; since each person's face has uniqueness and irreproducibility, the face recognition technology based on the human face has security and science, and it is being more and more widely used in our lives, bringing more security and convenience to our lives.
>>1、Pure ShadingLanguage draws HTML5 clocks2、VisualStudio uses three ways to launch the latest C features First method second method third method3、Leetcode14LongestCommonPrefix4、AI Minute The most soughtafter graduates are ranked Tsinghua did not make the top three Alipay to support iPhone X face swipe5、The 2013 Science and Research Foundation Awardwinning project milestones showcase is in full swing