Abstract: This paper presents an innovative face recognition method which is based on combination of graph isomorphism technique and Gabor wavelet. Here features are selected. Then algorithm determines the distance between these features in the face image. The algorithm applies Gabor wavelet on the selected features. Algorithm takes into account Gabor coefficient as well as Euclidean distance between features for face recognition. Brightness normalized and non normalized face images are given for face recognition to demonstrate that the normalized faces can improve the recognition rate.