Kinship Verification using algorithms in terms of appearance

Abstract

In recent decades, face verification and recognition issues have received considerable attention and are motivated by psychological results; several scientists have studied the ability of human assessors or the perception to recognize kinship from human facial images. In addition, the biological relationships and similarities between the traits found in the same family prompt us to exploit this fact to develop a computer system that will be able to recognize kinship by proposition an algorithm that determines whether a pair of facial images belonging to a class is '' kin '' or '' non. Our approach is based on the processing which ensures the preparation of the presentation of an image of the analysis. The latter is based on a step of extracting information in the important areas of the face of parents and children which are then compared between them between father and son, between father and daughter, between mother and son or between a mother and daughter this applies different algorithms in terms of type of appearance on the Kinfacew-II database. In the end, we tried logically to compare these methodologies one of them can provide the most accurate results for our hypothesis, so we found that the PCA algorithm gave the best estimate of 71, 05%. This result is encouraging and promising, as recent studies have concluded in this regard.