APPLICATION OF FRACTAL PROCESSING OF DIGITAL MAMMOGRAMS IN DESIGNING DECISION SUPPORT SYSTEMS IN MEDICINE
Article Sidebar
Main Article Content
Abstract
The subject of the research is the methods of digital processing of halftone medical images with locally concentrated features. The object of the research is the process of morphological analysis of digital mammograms in the design of decision support systems in medicine. The aim of this work is to develop methods and technologies for detecting of diagnostically significant characteristics of digital mammograms based on their morphological analysis, taking into account fractal dimensions. The objective of the study is to improve the quality of mammographic examinations of patients in the design of decision support systems in medicine by developing specialized methods for morphological analysis of digital mammograms (highlighting diagnostically significant elements amid noises), based on taking into account the features of the images in the form of useful signal models, in particular, fractal dimension models. Research methods: a method for calculating the fractal dimension of two-dimensional halftone images specified on a discrete set, methods of object-oriented programming, methods of statistical analysis. As a result of the research, the following results were obtained: based on the analysis of known methods of digital image processing, a limited area of their application in processing mammograms was shown and the urgency of developing specialized methods of morphological analysis based on taking into account the features of the considered images in the form of useful signal models, in particular, fractal dimension models. A method and an algorithm for the implementation of morphological analysis of digital mammograms, taking into account their fractal dimension, have been developed. The software implementation of the method was performed using the MatLab math package and testing on real mammograms was completed. Mammograms without obvious pathologies and mammograms which having pathological structures of various types (tumors, intraductal formations and microcalcifications) were processed. The fractal dimension of the entire image and selected fragments was calculated. Conclusions. The results of the research showed that the fractal dimension of the entire image does not give statistically significant results on the presence or absence of pathologies, but if we calculate the fractal dimension on the selected fragments, then the results are very different. We can trace the pattern that the more obvious pathologies on a fragment, the greater the fractal dimension. Further research is aimed at developing a method for classifying digital mammograms taking into account their fractal dimensions.
Article Details
References
Avrunin, O.Gh. and Bodjansjkyj, Je.V. (2018), Modern intellectual technologies of functional medical diagnostics: monograph, KhNURE, Kharkiv, 236 p. (in Ukrainian).
Povoroznyuk, A. I. (2011), Decision support systems in medical diagnostics. Synthesis of structured models and decision rules, LAP LAMBERT Academic Publishing GmbH & Co. KG, Saarbrücken Germany, 314 p. (in Russian).
Povoroznyuk, A.I., Povoroznyuk, O.A. and Mumladze, G. R. (2019), “Development of a functional model of the stages of diagnostic and therapeutic measures in decision support systems in medicine”, Control, navigation and communication systems, Issue 2 (54), PNTU, Poltava, pp.144-148 (in Russian).
Povoroznyuk, A.I. and Filatova, A.E. (2019), “Formalization of the problem of the matched morphological filtering of biomedical signals and images”, Information Technology in Medical Diagnostics II: Proc. of the International Scientific, Internet Conf. «Computer Graphics and Image Processing» and the XLVIIIth International Scientific and Practical Conf. «Application of Lasers in Medicine and Biology», CRC Press, London, UK, pp. 155-162.
Guo, Q., Ruiz, V., Shao, J. and Guo, F. (2005), “A novel approach to mass abnormality detection in mammographic images”, Proceedings of the IASTED International Conference on Biomedical Engineering, Innsbruck, pp. 180-185.
Povoroznyuk, A.I., Filatova, A.E. and Kozak, L.M. (2017), “Grayscale morphological filter based on local statistics”, The International Society for Optical Engineering, Vol. 10445, pp. 205-214.
Povoroznyuk, A.I., Filatova, A.E. and Shehna, Kh. (2019), “Development of Method of Matched Morphological Filtering of Biomedical Signals and Images”, Automatic Control and Computer Sciences, Vol. 53, No. 3, pp. 253-262.
Rubis, A.Yu., Lebedev, M.A., Vizil'ter, Yu.V. and Vyglov, O.V. (2016), “Morphological filtering of images based on cross-contrast”, Computer optics, Vol. 40, No. 1, pp. 73–79 (in Russian).
Yu.V. Vizilter, V.S. Gorbatsevich, A.YU. Rubis, O.V. Vygolov (2015) “Comparison of images by shape using diffuse morphology and diffuse correlation”, Computer optics, Vol. 39, No. 2, pp. 265–274 (in Russian).
Ajay Kumar, Bisoi and Jibitesh, Mishra (2001), “On calculation of fractal dimension of images”, Pattern Recognition Letters, Vol. 22, Issues 6-7, pp. 631-637.
Ampilova, N.В., Soloviev, I.Р. and Shupletsov, Yu.V. (2014), “On Fractal, Statistical and Morphological Methods of Digital Image Analysis in Medical Research”, Computer Science. Telecommunications and Control Systems, St. Petersburg State Polytechnical University Journal, No. 1 (188), pp. 51-61.
Povoroznyuk, A.I. and Shekhna, K.H. (2018), “Fractal processing of digital medical images in the design of decision support systems in medicine”, Information technology and computer modeling; materials of articles of the International scientific-practical conference, Ivano-Frankivsk, May 14-19, Ivano-Frankivsk, Mr. Goliney OM, pp. 77-80 (in Ukrainian).
Dyakonov, V.P. (2012), MATLAB. Complete tutorial, DMK-Press, Moscow, 768 p. (in Russian).