Purpose. Improving the efficiency of the objects identification computational methods using their monochrome images through the development of efficient technologies, methods and means of processing of digital representations (signals) with variable and changing information entropy. Development of the object identification method based on processing of probabilistic representations and projections that are formed using statistical evaluation of information entropy values of related fragments of objects characteristics. Investigation of the efficiency of correlation based processing of signal entropy representations and projections that are implemented based on the proposed approach. Methodology. The main scientific results and conclusions were made on a basis of information theory, theory of signals and systems, probability theory and mathematical statistics, methods of spectral and statistical analysis, digital signal processing and based on results of computational modeling. Findings. Based on computational modeling it was discovered that the proposed computational method allows achieving of unambiguous identification under information loss caused by image distortion that does not exceed 87 %. Besides, the independence of the proposed approach from inverse representation in the case of monochrome images was confirmed. Originality. The improved computational projection method of the object identification for which construction of projections is proposed to be done by statistical evaluation of information entropy values of corresponding fragments of 2D binary matrices of objects monochrome representations. Practical value. The proposed method allows increasing the correctness of identification in comparison with known projection based computational methods.

  Якщо, ви не знайшли інформацію про автора(ів) публікації, маєте бажання виправити або відобразити більш докладну інформацію про науковців України запрошуємо заповнити "Анкету науковця"