Geek asks and answers

Suggestions of algorithms for distinguishing images of human faces

I am looking to develop a face algorithm that will take 2 input images and calculates whether the faces in the two images will be the same person or not. I have all the image preprocessing working using OpenCV. In the prepossessing steps, I crop the face from the whole image, convert it to gray and resize it to a standard size. My problem is slightly different from face recognition in the sense that I am not interested in identifying a person in an image, I just want to say whether they are the same person or not. Most face recognition algorithms that I looked at rely on training on face samples of the same faces that will contain the input data. In my situation, I can train with common people, but not with the exact people who will be at the entrance.Could you suggest some methods that could work in my situation?

+3
source share
4 answers

For your problem, the perception of hashing along with the editing distance may be more correct.

Departure

+1
source

Usually you measure the distance between functions.

If you cannot control the position / size / angle of the image - that is, the image of the real world, and not someone sitting in front of the face scanner - then you need to think about how to make all these relationships.

Try to find the eyes (lungs), mouth or nose and get the sides / angle of the triangle. Then you can also make a percentage of the area of ​​the face covered with hair, the ratio of the length and length of the nose, etc.

+1
source

:

+1

If you haven’t done so already, check out Wavelet’s methods. See this document for example.

0
source

More articles:


All Articles