Abstract: One of the urgent and rapidly developing areas of modern information technology is related to the tasks of identifying an individual on the biometric characteristics of a person. Advantages of personal identification systems on the base of face images are its unobtrusiveness, passivity and relatively low cost. Despite this, the problems of extracting the features of a face image have not been studied enough. The feature extraction stage, which is the main step in the creation of systems for identifying a person's identity on these features basis, is considered. To solve this problem, an algorithm for extracting the features of the face images, consisting of the following stages, is proposed: detection of a face area on an image; search pupils on the face image; determining the location of the mouth and nose; formation of geometric features; extraction of subsets of strongly correlated features; determining representative feature in each subset of strongly correlated features. A distinctive feature of the proposed algorithm is the integration of the results of various methods for determining anthropometric points. The main purpose of this article is to develop an algorithm for identifying geometric features when identifying a person from a face image. The subject of the study are algorithms for preprocessing of a face image. In scientific terms, the results of this work together constitute a new solution to the scientific problem associated with the preprocessing of an image of a person in identifying a person's identity. The practical significance of the results lies in the fact that the algorithms and programs developed can be applied in forensics, access control, video surveillance, and security of information systems. To test the efficiency of the proposed algorithm, an experimental study was carried out to solve the problem of identifying a person's identity from a face image.
Index terms: face localization, facial feature points extraction, person identification.