Analysis of the principle of face recognition technology

Do you remember these plots in the movie? In "Transformers 2", when the young actor and his friends passed the checkpoint, they were deceived by the military on duty, but they were discovered by the military's face recognition technology. In the 2014 remake of the "Mechanical Men", the mechanical warfare was first publicly exposed to the public, and all the faces were scanned in the crowd, and the acquired faces were compared in the wanted criminal database. Yes, it was discovered in an instant that there was a wanted man who had escaped for many years and subdued it. There are many other movies, but all the secret departments of the United States will scan a variety of biometric features, from the fingerprints and irises of the early movies to the current faces.

What is face recognition?

Face recognition is a subdivision problem of visual pattern recognition, and it is probably the most difficult problem to solve. In fact, we are doing visual pattern recognition all the time, we obtain visual information through the eyes, and this information is recognized as a meaningful concept through the processing of the brain. So we know that there is a cup, a book, or something else in front of us.

We also do face recognition all the time. We encounter countless people in our daily lives, recognize those acquaintances, greet them, deal with them, and ignore other strangers. Even avoid those who owe money and are still not available.

However, this seemingly simple task is not so easy to implement for the machine.

For a computer, an image of a picture, whether it is a static picture or a frame in a dynamic video, is a matrix of pixels. For example, a 1080p digital image is a matrix consisting of 1980*1080 pixels. Each pixel, if it is an 8-bit rgb format, is three numbers ranging from 0-255.

The machine needs to find out in a certain part of the data which part of the data represents the concept: which part of the data is the cup, which part is the book, and which part is the face, which is the rough classification problem in visual pattern recognition.

Face recognition, in the part of the data that all machines consider to be human faces, distinguishes who this face belongs to. This is a fine classification problem.

How many types of faces can be divided into?

Depending on the size of the face database of the problem being processed, how many target faces are in the face database, the machine needs to perform a corresponding number of fine classification. If you want the machine to recognize everyone he sees, how many people in the world, how many faces can be divided into categories, and the difference between these categories is very subtle. This shows the difficulty of the face recognition problem.

Not to mention, this matter is also affected by various factors such as lighting, angles, and decorative objects on the face. It is not difficult to explain why face recognition technology is not widely used in daily life. Most people can only contact face recognition in science fiction movies.

Silly and unclear - some concepts that are easily confused

Some things that are less familiar are often accompanied by a lot of conceptual confusion.

For example, Chinese people who don't know much about Western religion may not know why some people believe in God but don't believe in Jesus. They are all uncles working in churches. Why do they want to abstain from sex and some can marry?

As a new thing, face recognition is accompanied by a lot of concept confusion, and distinguishing these concepts is still important for understanding face recognition.

Face Shield

Face Shield,Safety Face Shield,Medical Face Shield,Protective Face Shield

Ningbo Anbo United Electric Appliance Co.,ltd , https://www.airfryerfactory.com