The process of recognizing objects (in our case, the faces) is usually efficient if it is based on the acquisition of functions that include additional information about the class of object to be taken. In this tutorial, we will use the Haar-like features and the local binary patterns (LBP) to encode the contras ts highlighted by the human face and its spatial relationships with the other objects present in the image. Usually, these features are extracted using a cascade classifier that must be trained to accurately recognize different objects: the classification of faces will be very different from the classification of the car.