Open AccessДІАГНОСТИКА ЗАХВОРЮВАНЬ КРОВІ НА ОСНОВІ КЛАСТЕРНОГО АНАЛІЗУ ЗОБРАЖЕНЬ ГАЗОРОЗРЯДНОГО СВІТІННЯ
Author(s) -
N. V. Hlukhova,
L. A. Pisotska
Publication year - 2015
Publication title -
medična ìnformatika ta ìnženerìâ
Language(s) - English
Resource type - Journals
eISSN - 1997-7468
pISSN - 1996-1960
DOI - 10.11603/mie.1996-1960.2014.3.3808
Subject(s) - histogram , computer science , reliability (semiconductor) , fuzzy logic , pixel , artificial intelligence , cluster (spacecraft) , data mining , medical physics , computer vision , pattern recognition (psychology) , reliability engineering , medicine , image (mathematics) , engineering , power (physics) , physics , quantum mechanics , programming language
The problems of diagnosis of diseases of blood and the risk of their development, the analysis of the shortcomings of existing methods. A new approach to the implementation of screening studies, which is based on image capture gasdischarge radiation smears or drops of blood in the electromagnetic field of high voltage on x-ray film. In order to improve the reliability of diagnostic decisions by eliminating subjective factors proposed automated computer processing of images glow blood-based histogram brightness of pixels and performing procedures fuzzy cluster analysis method «fuzzy c-means». Informed choice of methodology radiation imaging of biological objects yielded satisfactory results in the field of diagnostic screening for blood diseases, early definition of risk incurred. The practical value of the proposed diagnostic technique is confirmed by the results of experimental data using cluster analysis.