PENGENALAN WAJAH DENGAN MATRIKS KOOKURENSI ARAS KEABUAN DAN JARINGAN SYARAF TIRUAN PROBABILISTIK

Abstract

Face recognition system is a development of the basic methods of authentication system using the natural characteristics of the human face as a baseline. Face recognition process consists of several phases, training and testing phase. The testing phase is done directly and indirectly. Indirect data test taken from a set of face images that have been selected, while direct data test take face image from camera. Human face recognition combines GLCM and PNN methods. Preprocessing is done by converting RGB to grayscale, using centroid method as face image segmentation process. Face recognition includes some factors, i.e. lighting, distance, angle and position. GLCM uses statistic method and second-order texture analysis, which represents image texture in following parameters energy, corelation, homogenity and contrast. While PNN is used to build database which is stored in the network in order to compare outcome from GLM in the form of matrix. This research uses face image as database by collecting sample of 10 persons, 5 face position, 2 type of distance shooting and 3 type of lighting. Testing process results 92% in direct recognition and 93,33% in inderict recognition.