HUMAN FACE DETECTION IN COLORED IMAGES BASED ON GAUSSIAN FILTERS

  A. N asrabadi1,*, S. Azadi2, J. Haddad Nia3     1. Lecturer, Electrical Engineering Dep., Engineering Faculty, Tarbiat Moallem University of Sabzevar, Sabzevar, Iran   2. Assistant Professor , Electrical& Computer Engineering Faculty, Semnan University, Semnan, Iran   3. Assistant Professor , Electrical Engineering Dep., Engineering Faculty, Tarbiat Moallem University of Sabzevar, Sabzevar, Iran   * Corresponding Author : abbas.nasrabadi@gmail.com     Abstract   In this article a recommended method is introduced for face recognition in colored images, based on combination of template matching, conversions of color filter. In this method, firstly based on the recommended Gaussian filter, skin points of image based on a cluster including all the elements of skin are determined and by extraction of color elements, the standard of skin color separation is determined. Then the linked areas similar to skin color that are separated from different parts of image are introduced as recognized face by template matching and based on a model. The results of simulation on an induced images bank show that the above mentioned method has not only a low positive and negative mean error but also a high resistance against noise.     Keywords: Face Detection, Skin Model, YCbCr Colored Space, Gaussian Filter, Template Matching.