In the present investigation, the performance of a classifier based on facial analysis and detection of wrinkles is analyzed. The experimental part has been divided in two phases. The first, analyzed the proposed features, and the second, evaluated its performance on a database of faces. During the first phase, a statistical analysis of the sets analyzed has been carried out, and therefore, the correlation between the characteristics proposed for the project, have been studied and discussed. In the second phase, several tests -under different conditions- have been performed in order to determine the capability and efficiency of the system in relation to recognition tasks. In particular, the optimal size of the training set, and the performance of various measures of similarity of the two sets of proposed characteristics have been studied, also the subsets of discriminating characteristics have been evaluated.