Faces are familiar objects that can be easily perceived and recognized by ourselves.
However, the computational modeling of such apparently natural human ability remains
challenging. Recent studies in the literature have suggested that face processing is a
cognition task composed of configural (or global) and featural (or local) sources of information,
but with controversial arguments about the combination of these two types of
information. In this work, we describe an holistic method that combines variance used
in Principal Component Analysis (PCA) with some prior knowledge about the underlying
visual perception task, including systematically the global and local information
in the common multivariate representation of face images. We have showed that, with
prior information, important local variations represented by principal components with
small eigenvalues may not be discarded augmenting the classification accuracy of the first
orthogonal basis vectors. Most interestingly, PCA with prior knowledge provides a specialized
feature selection procedure, where the mapping of high-dimensional data into a
lower-dimensional space has been able to handle local variations and capture not only the
entire facial appearance but also some sample group facial features.