In this work, different algorithms and procedures used in the development of a system for the automatic extraction and classification of the lips postures, are presented. System was designed with the purpose of help to the children with lip and cleft palate in the recovery process after reconstruction surgery. For facial detection, techniques based on the YC bCr color space together with connectivity analysis, are used. In order to get a close model of the lip contour for features extraction, projection techniques, color analysis ( HSV and red exclusion) and SUSAN edge operator are used. For determining discriminant characteristics from the region described by the contour, different types of statistical analysis, are used. Classification of lips postures is made using different classifiers.