1
$\begingroup$

I am looking for best open source python repo for facial recognition. Best if it uses tensorflow backend. I know you can train images to recognize. Yolo can be used if trained on face. To name the person.

But I wonder if there is any code where you can add new faces to database without training or minimum training. As new faces are added I don't want to train the network repeatedly. Also the less amount of face picture needed the better.

If code is not available any guide or research paper will also be helpful. For example what approach can I take to make an app for a person who has difficulty remembering peoples name. So the app can take a small video or few photos with name and will be able to tell the persons name in the future. Neural network should not be retrained while adding new face to database if possible.

$\endgroup$
0
$\begingroup$

It seems your problem is more related to Face Identification than Face Recognition.

I understand you are looking for the implementation using a NN based approach, but if you're open to giving it a try to other approaches you could consider using Eigenfaces, which is based in PCA.

For that, you can find some references and code implementations.

Datasets you can use for testing face identification:

  • AT&T Face Database
  • Extended Yale Face Database B
  • BioID Face Database

References you can look at for additional implementation details: https://www.researchgate.net/publication/333367462_Improving_face_recognition_of_artificial_social_companions_for_smart_working_and_living_environments

A forked repository on GitHub for this approach: https://github.com/joaoquintas/facerec

Original repository for reference https://github.com/bytefish/facerec

$\endgroup$

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