Face Recognition by Optical Correlation using 3D Filter

Shape 3D recognition rose up recently a big interest. An increasing number of applications has been launched in this theme. In this paper we propose and describe a robust method based on a 3D filter concept to recognize a 3D color human face. This method coming from 3D image processing in optics image processing is investigated, namely 3D multidecision correlation for pattern recognition forms. For this technique and in order to approve decisions using a correlation technique, we propose an optimisation of correlation filters. This optimisation is carried out by taking into account spectrum phases of several references (several position) which are used to realize one 3D multidecision correlation filter