Deep Learning for Face Recognition. People find it hard to recognize people by their faces. They might not know what they look like.
It is a task that humans can do easily. It doesn’t matter what the light is like or if their faces are changed. Still, it has been hard for computers to do this well until recently.
Deep learning is a new way to learn information. It can use very large datasets of faces and learn new things about them. When you do this, it will know how the face works. This has allowed models that are able to identify people to do as well as humans, and now they can even do better.
Face recognition is when you try to find a person in a photo or video. If you can find them, then it is verified.
In the process of face recognition, you can first detect someone’s face. Next, align the person’s face with a picture of them so that it is easy to see. Then use some features like their nose and eyes for recognition.
Deep learning models are good at recognizing faces. They have even done better than humans!
Faces in Photographs
It can be hard to find someone in a photo. You might need to automatically recognize the people.
There are many reasons why we might want to automatically recognize a person in a photograph.
For example:
We might want to restrict access to a resource and only allow one person, called face authentication.
Or we might want to confirm that the person matches their ID. That is called face verification.
We may want to give a name to the person’s face. This is called face identification.
We refer to this as the problem of “automatic face recognition.” This can happen when a person is in a still photograph or in the video.
Can humans do better?
We can find the faces in pictures and tell who they are. Computers can do this very well, for example when people have aged, wear sunglasses, have different colored hair or look in different directions. We can even find faces where there are none, such as in clouds.
This is a difficult problem for computers to solve. Computers have been trying to solve this problem for a long time, but the problem is still not solved. Recently?
Process of Automatic Face Recognition
Face Recognition is the process of recognizing a person in a picture.
A machine can tell the difference between one person to another. They are looking for someone in a database of faces that looks like the person in the picture.
Face recognition is a hard thing to understand. It starts with face detection, which then goes to alignment. The next step is feature extraction, and finally the last step is face recognition.
Face Detection.
Find one or more people in the picture. Put a box around their face.
Face Alignment.
You need to make the face that is in the database match what you see. For example, if someone’s face is not square, then you need to adjust it.
Feature Extraction.
Extract features from the face for a recognition task.
Face Recognition.
Match a face with one or more faces in a database that you have prepared.
A system might have separate modules to handle each step, or it might combine some or all of the steps into a single process.
