Performing face recognition

Detecting faces is a fantastic feature of OpenCV and one that constitutes the basis for a more advanced operation: face recognition. What is face recognition? It is the ability of a program, given an image or a video feed containing a person's face, to identify that person. One of the ways to achieve this (and the approach adopted by OpenCV) is to train the program by feeding it a set of classified pictures (a facial database) and perform recognition based on the features of those pictures.

Another important feature of OpenCV's face recognition module is that each recognition has a confidence score, which allows us to set thresholds in real-life applications to limit the incidence of false identifications.

Let's start from the very beginning; to perform face recognition, we need faces to recognize. We fulfill this requirement in two ways: supply the images ourselves or obtain freely available face databases. A large directory of face databases...