What are the best ways to implement an attendance logging system from live video feed using face recognition?

I am working on a face recognition based attendance logging system. Since a live video feed contains multiple images of a same person, I wanted to know what are the best practices that can be used to recognise a person without any misclassification and delay?

[KALYAN ACHARJYA]

The concept is quite old it having challenges (Practically: Face fetch from live video or movement). Please look at the different FRA (Face Recognition Algorithms) , numerous literature and comparision avaliable. like PCA and CV approach.

Good Luck!

[Shaik Ahmad]

Dear @Kalyan

thank you for your inputs. I have started studying related FRAs. And I got a fair amount of knowledge on face recognition . But my concern is in live video feed, actually we get multiple images of a same person (let consider with a 30 fps camera if a person stays in the line of sight for 10 seconds I get 300 frames of the same person).

In this situation how do we calculate the final prediction in robust manner.