Face Recognition with Localization

Can We use localization to optimize the Face Recognition , I mean we can use localization to detect the faces in the image and then crop it , So We can get a photo for only faces without a Noise Background

Thank u

It could be possible to do it that way I think, unless speed is important to your application. This remind me of R-CNN model which has to propose regions of interest and then run again on those regions of interest to detect the objects, but its not as fast as YOLO.

In face recognition applications I would think speed is a key factor.