In order to get a feeling of how PCA could help us with our face model, we will start with an Active Shape Model and test some parameters.
Since face detection and tracking have been studied for a while, several face databases are available online for research purposes. We will use a couple of samples from the IMM database.
First, let's understand how the PCA class works in OpenCV. We can conclude from the documentation that the PCA class is used to compute a special basis for a set of vectors, which consist of eigenvectors of the covariance matrix computed from the input set of vectors. This class can also transform vectors to and from the new coordinate space, using project and backproject methods. This new coordinate ...
