Facial Feature Point Ratios are the third type of features whose effects on family resemblance are evaluated. These 9 ratios are calculated from the distances between the characteristic points of the face. In order to eliminate the dependence of the proposed algorithm on the image scale, this set of ratios is used instead of the distances between the feature points of the face. These ratios are as follows: Eleven distances used to calculate the above ratios are shown in Figure 6. Locating the exact coordinates of the facial feature points is crucial for calculating the ratios. As shown in Figure 5, the hairline, chin and sides of the face are points that form the boundaries of the face. These points are simultaneously located while the face is detected and cropped from the image. In order to extract other feature points from the face, two types of methods based on geometric features are used. Firstly, the Linear Principal Transform (LPT), proposed by Dehshibi et al [Deh10], is performed to locate the eyebrows, eyes, tip of the nose and centerline of the lips in frontal view of the face. Then, an extended version of LPT, which we called LPT2, is used to locate these points in the profile view. The Linear Principal Transform (LPT) is a one-to-one transformation, which has three key characteristics, including "precision," "power," and "simplicity." The main objective of LPT is to identify the most significant basis, which contains the features of interest. This will reveal the hidden structure of the data. LPT assumes that an m × n image consists of m sets of observations in an n-dimensional vector space. Among these vectors, the vector with the highest variance corresponds to the feature of interest. To obtain a feature, first, the covariance matrix of the image is calculated...... middle of paper ...... images are calculated. Then, based on the calculated weight, half of the training data is discarded. In the second step, database filtering is done using the “eye region”. In the last recognition step, the “frontal face” is used to find three images that have the minimum Euclidean distance to the input image. In order to evaluate the effectiveness of the proposed algorithm, a structure for the family must be considered. Compared to the FFIDB images, a three-level structure is defined. As shown in Figure 11, each level has an impact factor. The effectiveness rate of the proposed method is equal to (sum of the impact factor of each level)/(sum of the maximum impact factor). For example, if the images selected in the recognition phase have the “mother”, “sister” and “cousin” relationship with the input image, then the accuracy of the algorithm is 77,77 %..