ARAleksandr L. Reznik

Aleksandr L. Reznik

*This is an algorithmically generated author record based on publicly available information.

Aleksandr L. Reznik has a total of 26 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Software Support for Optimal Localization of Signal-Point Sources with Random Pulse Generation

    by Alexander L. Reznik, Alexander A. Soloviev

    The article analyzes algorithms for the fast-optimized dynamic search for point Poisson sources that generate ultrashort (instantaneous) delta pulses at random instants. A key distinguishing feature of the algorithms proposed in this article is that the authors developed and applied (and implemented on a specialized cluster) acceleration methods based on parallel computations to calculate the cumbersome multidimensional integral expressions that arise during the solution of the problem. An example of a software calculation of the parameters of an optimal search algorithm for several localized pulsed objects with a given localization accuracy is provided. The optimal number of scanning stages for minimizing the average localization time is calculated. The search strategies described in this paper are focused on localizing sources with a uniform distribution density, but the conducted research opens up further prospects for developing fast-optimized localization algorithms for random signal sources with arbitrary distribution densities.

  • Многомерное расширение классических чисел Каталана для решения непрерывных задач анализа случайных точечных изображений

    by А.л. Резник, А.а. Соловьев, И.ю. Резник, Р.м. Кожевников

    Предложено обобщение классических чисел Каталана на многомерный случай, базирующееся на словарно-символьной статистике. Представлены формулировки вероятностно-комбинаторных задач, решение которых приводит сначала к двумерному, а затем и трехмерному обобщению классической последовательности Каталана. Изложенный подход возник и вместе с программами компьютерной алгебры применялся при переводе в дискретную форму ряда непрерывных вероятностных задач, связанных с анализом случайных точечных изображений. Эффективность приведенного в работе многомерного расширения чисел Каталана продемонстрирована на примере решения с их помощью классической баллотировочной задачи Бертрана. Purpose. The purpose of the research is to systematize and find an explicit analytical form of generalized Catalan numbers based on a word-symbolic representation, as well as their application for applied problems related to the processing of random point images. Methodology. An original mechanism for reducing solvable continuous problems to their discrete analogues is proposed. In the practical implementation of the chosen research scheme, there are two types of problems, namely, discrete-combinatorial problems leading to generalized Catalan numbers and problems of calculating multidimensional integrals over convex polyhedra in n-dimensional space requiring the development of high-speed computer algebra programs. The main content. A generalization of the classical Catalan numbers to the multidimensional case is proposed, based on word-symbolic statistics using a limited alphabet. Probabilistic-combinatorial problems are explicitly formulated; their solution leads to a two-dimensional and then to a three dimensional generalization of the classical Catalan sequence. The main advantages of the new extension of Catalan numbers are the ease of generalization of word-symbolic problems to the multidimensional case and the unity and universality of their formulations. Findings. Word-symbolic problems leading to multidimensional generalized Catalan numbers are formulated. An explicit form of generalized Catalan numbers in two-dimensional and three dimensional cases is obtained. A number of probabilistic problems related to the analysis of random point images have been solved. Scientific novelty and originality. The originality and scientific novelty of the article contains both the obtained results and the developed methods to solve them. The article presents new previously unknown probabilistic formulas required for solving problems related to the registration and analysis of random point images. A package of programs has been created for high-speed integrating multidimensional integral expressions over areas limited by a system of hyperplanes in n-dimensional space.

  • Open Access

    Application of Analytical Methods and Computer Algebra Tools for Calculating Invariant Characteristics of Random Point Images

    by Александр Львович Резник, Alexander A. Soloviev, Александр Анатольевич Соловьев

    When solving a number of problems related to image processing, one of the most important points is the correct choice of features that determine the degree of “abnormality” of the image under study. The paper proposes a set of several classification features of random point images for their use in the tasks of detecting anomalous clumps in the processed data stream or, on the contrary, detecting areas and fragments with increased inter-element spread. Numerical values of features are calculated over the ensemble of analyzed images. Their significant difference from the pre-calculated theoretical values that correspond to random point images can help identify the presence of a regular component or an “abnormal” component in the processed data array. Theoretical and software calculations performed using specialized computer algebra tools are presented in the paper. The discussed classification features and probabilistic dependencies are stable invariants characterizing a random point image.

  • Open Access

    Algorithms for Increasing the Resolution of Digital Images Using Circular Scanning Data

    by Александр Львович Резник, Alexander A. Soloviev, Александр Анатольевич Соловьев

    The article describes an algorithm for increasing the spatial resolution of digital images using circular scanning technology. A methodology for solving the entire chain of problems associated with circular scanning of images is described in full details. It starts with the development of the necessary mathematical model and ends with the final programs and algorithms for optimal subpixel reconstruction of original digital images with increased spatial resolution. The optimal values of the scanning radius are calculated using the software modelling method. It means calculating the relative radius of the photomatrix circular trajectory with the minimal root-mean-square error when reconstructing the original noisy image. Examples of various reconstructed images are shown along with the calculations of reconstruction accuracy for various parameters of circular scanning and levels of multiplicative and additive noise. Software implementation of the developed algorithms is done using the MATLAB system.

  • Software-Analytical Calculation of Invariant Characteristics of Random Point Images Based on Order Statistics

    by A. L. Reznik, A. A. Soloviev

    A set of classification features of random point images is proposed for their use in problems of identifying anomalous concentrations in a processed data stream or, conversely, identifying areas and fragments with increased inter-element dispersion. A significant deviation of the numerical values of features calculated from the ensemble of analyzed experimental images from pre-calculated theoretical values corresponding to random point images may indicate the presence of a regular component or an “anomalous” component in the processed data array. Theoretical and software calculations performed using specialized computer algebra tools are presented.


  • Comparative Analysis Multistage Algorithms for the Localization of Random Pulsed-Point Sources by Efficiency

    by A. L. Reznik, A. A. Solov’EV

    Different multistage procedures for the localization of pulsed-point sources manifested by generating instantaneous pulses at random time moments are compa


  • Solving Fundamental and Applied Problems of Digital Image Processing at the Institute of Automation and Electrometry and Other Scientific Schools of the Siberian Branch of the Russian Academy of Sciences

    by A. L. Reznik, A. A. Soloviev

    The most important results obtained in recent years by the research institutes of the Siberian Branch of the Russian Academy of Sciences in the field of development of mathematical methods and the construction of effective information and computing systems for solving fundamental and applied problems of digital image processing are presented. Examples are given of the development of specific high-performance software and hardware systems intended to effectively solve important theoretical and scientific-applied problems, the solutions to which is based on the use of advanced methods of digital image processing.

  • Methods, algorithms and programs of computer algebra in problems of registration and analysis of random point structures

    by A.L. Reznik, Institute of Automation and Electrometry of the Siberian Branch of the Russian Academy of Sciences, A.A. Soloviev


  • Open Access

    Solving the Applied Problem of Random-Dot Images Analysis Using a Multidimensional Extension of Classic Catalan Numbers

    by Александр Львович Резник, A.A. Soloviev

    Введенное в статье понятие обобщенной последовательности Каталана является полезным инструментом при решении многих теоретических и прикладных вероятностно-комбинаторных задач. В сочетании с алгоритмами, осуществляющими программные аналитические преобразования, обобщенные числа Каталана упрощают решение многих задач информатики и прикладной математики. В частности, они оказываются эффективным средством решения задач, относящихся к регистрации случайных точечных изображений, при построении преобразований сигналов различной степени гладкости, при разработке оптимальных по быстродействию алгоритмов поиска импульсно-точечных объектов со случайным временем генерации мгновенных импульсов. Предложение авторов формулировать задачи перечислительной комбинаторики в словарно-символьной форме естественным образом приводит к многомерным расширениям классических чисел Каталана и обладает несколькими преимуществами. Совместное применение многомерных чисел Каталана и высокопроизводительных систем компьютерной алгебры позволило авторам решить ряд сложных прикладных задач, связанных с надежностью регистрации случайных точечных изображений.

  • Software and Combinatorial-Probabilistic Tools for the Analysis of Random Point Structures

    by A. L. Reznik, A. A. Soloviev

    Approaches to solving problems of processing random point images are described, which are based on the creation of high-performance software systems for carrying out labor-intensive analytical transformations. Original methods for calculating probabilistic relations expressed in the form of multidimensional integral expressions over convex polyhedra in n-dimensional space are implemented in software. A model for transferring the continuous problem of estimating the probability of error-free reading of a random point image into the class of discrete-combinatorial ones is proposed. Using the symbolic-geometric approach, an explicit form of generalized multidimensional Catalan numbers is found.