Algorithm for Selecting Optimal Features in Face Recognition Systems

Biometric methods are divided into methods based on physical and habitual movements. Physical methods are based on the anatomical uniqueness of each person. Anatomical features include facial features, eye pupils, fingerprints, hand geometry, ear lobes, and other biometric features. Face recognition systems are based on anatomical features, the quality of which is related to the efficiency of feature extraction. Methods for generating face image features are usually based on the separation of local and global features. Algorithms for extracting local features extract individual parts of the face image and perform recognition based on them. Algorithms for extracting global features perform the formation of features on the entire face image. At the same time, the recognition of these features based on optimality allows to reduce costs. This article is devoted to the analysis of algorithms for the formation of the space of optimal features of a face image, which proposes the task of identifying optimal features, as well as a method and algorithm for extracting optimal features of a face image.