The article deals with one of the key problems of the social network analysis – the problem of classifying accounts based on media content uploaded by users. The main difficulties are the content heterogeneity (both in format and subject) and the large volumes of data, which leads to excessive computational complexity of its processing and often to the complete inefficiency of traditional analysis methods. In the article, we discuss an approach to the clustering of media content from social networks based on textual annotation using BigData technology – a modern and efficient tool that allows to solve the problem of large data volume processing. To carry out computational experiments, a large sample of heterogeneous images (photographs, paintings, postcards, etc.) was collected from real Twitter accounts. The results confirmed the high quality of media content clustering, the average error was around 5 %.

In this work, technologies of analog optical computing for image edge detection using flat interfaces between dielectrics are studied both theoretically and experimentally. The feasibility of analog differentiation and edge detection by using the Brewster effect for polarized light sources is demonstrated. The efficiency of the analog image processing is investigated using a variety of amplitude and phase masks.

The paper presents an analysis of various approaches to constructing descriptions for the gradient fields of digital images. The analyzed approaches are based on the well-known methods for data dimensionality reduction, such as Principal (PCA) and Independent (ICA) Component Analysis, Linear Discriminant Analysis (LDA). We apply these methods not to the original image, represented as a two-dimensional field of brightness (a halftone image), but to its secondary representation in the form of a two-dimensional gradient field, that is, a complex-valued image. In this case, approaches based on using both the entire gradient field and only its phase component are considered. In addition, two independent ways of forming the final description of the original object are considered: using expansion coefficients of the gradient field in a derived basis and using an original authors' design that is called model-oriented descriptors. With the latter, the number of real coefficients used in the description of the original object can be halved. The studies are conducted via solving a face recognition problem. The effectiveness of the analyzed methods is demonstrated by applying them to images from Extended Yale Face Database B. The comparison is made using a nearest neighbor's classifier.

The work continues the author's previous study of positional number systems in finite fields. The paper considers ternary number systems and arithmetic operations algorithms for the representation of elements of finite fields in the so-called ternary reduced number systems, which are reductions of the canonical number systems when mapping the corresponding ring of integers of a quadratic field into some prime field. A classification of finite fields in which such number systems exist is given. It is proved that the reduced ternary number systems exist for most finite prime fields.

In this paper we synthesized and measured optical characteristics of an azopolymer based on the polymer N-epoxypropylcarbazole and chromophore 4 - (4-nytrophenylazo)-aniline. Regularities in surface microstructure formation under the influence of a Gaussian beam focused on the synthesized polymer film were shown. With the linearly polarized incident laser beam, the anisotropic formation of the surface microstructures was demonstrated. The dependence of the microstructure formation on the incident beam power density was studied. Also, we experimentally investigated the nonlinear optical effects emerging in the azopolymer structure exposed to a high-intensity laser beam. Nonlinear topographic effects arising during the formation of microstructures by means of linearly and elliptically polarized beams were described.