GRAYSCALE IMAGES IMPROVEMENT BASED ON AUTOMATED BRIGHTNESS ESTIMATION OF FUZZY MEMBERSHIP FUNCTIONS
DOI:
https://doi.org/10.32782/KNTU2618-0340/2021.4.1.2Keywords:
image processing, fuzzy membership function, fuzzy sets of type_2, singular value decompositionAbstract
Images that were generated by various systems, which are the result of standard research methods, often have insufficient quality for reliable analysis. They contain distortions caused both by the system of their formation and by the methods of presentation and showing in the processing system. To increase the reliability of image analysis, it is necessary to improve their visual characteristics in terms of identifying objects of interest for solving a specific problem. The modern approach to solving the problem of image analysis due to inaccuracy, incompleteness of initial data and ambiguity of processing algorithms (for example, when determining classes, regions/boundaries of objects) is based on the usage of fuzzy methods. This paper considers the information capabilities of characteristics synthesized on the basis of the method of singular value decomposition in a fuzzy feature space for improving the quality of grayscale images. The existing approaches to the usage of fuzzy functions of type_2 and the influence of the method of their formation on the result are described. The proposed algorithm uses the statistical characteristics of fuzzy membership functions of type_1 for transition to fuzzy sets of type_2. Proposed in this work method allows the automated selection of the most informative fuzzy components, based on the analysis of their brightness characteristics, at the stage of defuzzification using singular decomposition. The algorithm of proposed method and experimental results are presented on the example of a real microscopic image for various methods of preprocessing of the initial data, which demonstrate that the preprocessing of the initial data significantly affects the sensitivity of the transformation. It is shown that the transition to a fuzzy space of type_2 features, followed by the usage of a singular transformation with a preliminary selection of the most informative fuzzy membership functions, which are interpreted as images, based on an automated estimation of their brightness, provides an improvement in the visual characteristics of grayscale images.
References
Bezdek J.C., Keller J., Krishnapuram R., Pal N.R. Fuzzy Models and Algorithms for Pattern Recognition and Image Processing. Handbooks of Fuzzy Sets series. Boston: Kluwer Academic Publisher 1999. 678 p.
Tizhoosh H.R., HauBecker H. Fuzzy Image Processing: An Overview. Handbook on Computer Vision and Applications. Academic Press. 1999. Vol. 2. P. 683–727.
Fuzzy Sets and Their Extensions: Representation, Aggregation and Models / edited by Bustince H., Herrera F., Montero J. Springer, 2008. 674 p.
Bloch I. Signal and Image Processing. Telecom ParisTech, 2008. 295 p.
Handbook of Biomedical Imaging / edited by Paragios N., Duncan J., Ayache N. Springer, 2015. 308 p.
Zadeh L.A. The concept of a linguistic variable and its application to approximate reasoning. Information Sciences. 1975. Vol. 8. P. 199–249.
Сhi Z., Yan H., Pham T. Fuzzy algorithms: With Applications to Image Processing and Pattern Recognition. Singapore; – New Jersey; – London; – Hong Kong : Word Scientific, 1998. 225 p.
Castillo O., Melin P. Type-2 Fuzzy Logic: Theory and Applications. Springer-Verlag, 2008. 223 p.
Mendel J.M., John R. Type 2 Fuzzy Sets Made Simple. IEEE Transactions On Fuzzy Systems. 2002. Vol. 10. No 2. P. 117–127.
Mendel J.M., Robert I.J., Feilong L. Interval Type 2 Fuzzy Logic Systems Made Simple. IEEE Transactions on Fuzzy Systems. 2006. Vol. 14. No 6. P. 808–821. https://doi.org/10.32782/KNTU2618-0340/2021.4.1.2
Akhmetshina L., Yegorov A. Iprovement of Grayscale Images in Orthogonal Basis of the Type‐2 Membership Function. CMIS-2021: The Fourth International Workshop on Computer Modeling and Intelligent Systems, Zaporizhzhia, April 27 2021. P. 465–474.
