Facial recognition is a way of recognizing a human face through technology. A facial recognition system uses biometrics to map facial features from a photograph or video
Refer the following link. It may be useful for you.
I recommend reviewing Labeled Faces in the Wild LFW dataset results:
http://vis-www.cs.umass.edu/lfw/results.html
This is a challenging dataset, in which the latest state-of-the-art algorithms and patents are tested. A feature of this dataset is that it presents several scenarios of interest in facial recognition systems such as:
Unsupervised
Image-Restricted, No Outside Data
Unrestricted, No Outside Data
Image-Restricted, Label-Free Outside Data
Unrestricted, Label-Free Outside Data
Unrestricted, Labeled Outside Data
Human Performance
I have studied this problem for some years, and I have also implemented some facial recognition algorithms from scratch and from my knowledge, these are some pending algorithms challenges:
Difficulty working with large databases
Invariance to illumination
Need to retrain
Pose and expression changes well handled with key point detection algorithms, for example: https://ieeexplore.ieee.org/abstract/document/6296663