Family of recognition algorithms based on the selection of two-dimensional representative pseudo-objects in the training set

Abstract

Recognizing objects within large-dimensional feature spaces presents significant challenges in constructing effective recognition algorithms. This study proposes a groundbreaking approach that harnesses two-dimensional threshold functions to address the recognition problem with precision. The key innovation lies in the selection of representative pseudo-objects, which serve as the foundation for constructing two-dimensional threshold functions in the recognition algorithm model. Experimental studies were conducted, focusing specifically on face recognition, to assess the performance of these algorithms. The results of the above studies have shown that the proposed statistical algorithms increase the recognition accuracy and reduce the time of recognition of face parts and the face as a whole, described in the space of interrelated features. This study highlights the potential applications of these algorithms in diverse software systems tailored to solving practical recognition problems within the constraints of large-dimensional feature spaces.