0
$\begingroup$

I need to use an accurate library for face recognition in my attendance python project.

I used the face-recognition but I could no get the desired result. Is there any way to get 99% accuracy? I have more than 1000 students to register and then need to recognise them.

$\endgroup$
1
  • $\begingroup$ What results are you getting? $\endgroup$ Nov 20, 2023 at 13:42

1 Answer 1

1
$\begingroup$

First of all, I want to recommend this article (TABLE IV). It's not the most recent, but the algorithm families haven't changed much.

If you were referring to the face_recognition library, then it uses the dlib_face_recognition_resnet_model_v1 model, which is indeed somewhat outdated.

In my opinion, you should compare the two models Facenet and SphereFace2. The second one is quite new. And for training the first one, Google used a huge amount of data (more than for any other model).

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.