The classical approach is about choosing features using domain knowledge of the data to create features, which are then classified through a machine learning algorithm.
The neural network will find features itself. This works on large data sets and is invariant to pose, illuminations, etc. Facebook’s DeepFace and Google’s FaceNet use this approach.
Histogram of Oriented Gradients (HOG)
This technique can spot image gradient or intensity change in localized portions of the image to extract features related to the edges and shapes. HOG features are classified with a Support Vector Machine classifier for face detection.