• Nataliya Blavatska UIAZ Department of the Center for Strategic Communications of the Educational and Scientific Institute of Information Security and Strategic Communications of the National Academy of Security of Ukraine
  • Valeriy Kozura Department of Technical and Scientific Research of the Cyber Security Center of the Educational and Scientific Institute of Information Security and Strategic Communications of the National Academy of Security of Ukraine



recognition systems, approximation, correlations, binary features


The article considers the problem of object recognition by features in the process of deep learning and proposes a method of approximation of the multidimensional discrete probability distribution of features for efficient use of device memory. To achieve high recognition accuracy, a unified approach is used in the work, which provides an adequate balance with the accuracy of the results while reducing the amount of memory necessary for storing reference objects. The authors of the article consider the importance of considering the correlations between the features of objects, which contribute to increasing the efficiency of the recognition system. They show that computing probability distributions based on a limited number of parameters can significantly reduce the amount of training data needed to establish class standards for recognition. The results of the work emphasize that the complex approximation method can be successfully applied on various types of computers, including personal computers and specialized digital devices. The results of this study are important in the context of the development and optimization of such systems, as they are aimed at improving object recognition in deep learning systems under conditions of limited memory and data resources.


Ито Т., Фукусима М., Судзуки К. Статистическая тео¬рия распознавания образов. «Мицубиси дэнки гихо», 1973, т. 47 № 8, С. 881-892.

Fukunaga K. Introduction to statistical pattern recognition. Second Edition. Academic Press, Inc., Boston, San Diego, New York, London, 1990. 591 p.

Duda R.O., Hart P.E., Stork D.G. Pattern classification (2nd edition), Wiley-Interscience, New York, 2001, 738 p.

Schlesinger M.I., Hlaváč V. Ten Lectures on Statistical and Structural Pattern Recognition // Computa-tional Imaging and Vision (CIVI), vol. 24. Kluwer Academic Publishers. Dordrecht / Boston / London, 2002. 519 p.

Ivakhnenko A.G., Ivakhnenko G.A. The Review of Problems Solvable by Algorithms of the Group Method of Data Handling (GMDH), Pattern Recognition and Image Analysis, vol.5, 4, 1995. pp. 527-535.

Bahadur R.R. A representation of the joint distribu-tion of responses to n dichotomous items, in Studies in Item Analysis and Prediction, pp. 158-168, H. Solomon, ed. (Stanford University Press, Calif., 1961).

Chow C.K., Liu C.N. Approximating discrete proba-bility distributions with dependence trees, IEEE Trans. Info. Theory, IT-14, pp. 462-467 (May 1968).