МУЛЬТИФРАКТАЛЬНИЙ АНАЛІЗ МАМОГРАФІЧНИХ ЗОБРАЖЕНЬ
Системи управління, навігації та зв’язку. Збірник наукових праць · 04.05.2026 · 2 (84) · 129-133
DOI: 10.26906/sunz.2026.2.129 ↗Kurz gefasst
Fachartikel in der Zeitschrift Системи управління, навігації та зв’язку. Збірник наукових праць, online erschienen am 04.05.2026. Verfasst von Oleksandr Lychkatyi und 1 weiteren Autor:innen. Die Arbeit nennt 8 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0).
Abstract (Originalsprache)
The study addresses the problem of improving the effectiveness of computer-aided analysis of mammographic images to support the early detection of pathological changes in breast tissue. The limitations of traditional mammography—caused by tissue density, the presence of noise and imaging artifacts, as well as the subjectivity of visual interpretation—justify the need for quantitative and objective analytical methods capable of providing a more reliable assessment of structural features. The work examines an approach aimed at obtaining locally sensitive indicators of tissue complexity, which enhances the detectability of low-contrast or weakly expressed abnormalities. The objective of the research is to develop a methodology that combines automated segmentation with multifractal analysis and presents the results as maps of local fractal dimensions (heatmaps). Segmentation is performed by combining a U-Net neural network with Otsu thresholding after median and Gaussian filtering, which produces more stable masks and reduces the influence of noise components. A key element of the proposed methodology is the construction of local fractal-dimension maps using a sliding-window technique, which enables continuous spatial assessment of structural heterogeneity. It is shown that regions containing pathological changes exhibit localized deviations of the fractal dimension relative to the surrounding parenchyma, thereby improving interpretability and potentially indicating areas of interest. The study analyzes the limitations of the classical box-counting method when applied to grayscale images, particularly its dependence on binarization and the loss of intensity information. The rationale for employing differential box-counting is presented, as this method operates in the intensity domain and provides greater sensitivity to subtle textural variations. Experimental results obtained using images from the MIAS and DDSM datasets confirm the potential of the proposed approach and highlight its usefulness for further development of computer-aided decision-support systems in mammography.
Zitieren
APA
Lychkatyi, O., & Povoroznyuk, A. (2026). МУЛЬТИФРАКТАЛЬНИЙ АНАЛІЗ МАМОГРАФІЧНИХ ЗОБРАЖЕНЬ. Системи управління, навігації та зв’язку. Збірник наукових праць, 2(84), 129-133. https://doi.org/10.26906/sunz.2026.2.129
Vancouver
Lychkatyi O, Povoroznyuk A. МУЛЬТИФРАКТАЛЬНИЙ АНАЛІЗ МАМОГРАФІЧНИХ ЗОБРАЖЕНЬ. Системи управління, навігації та зв’язку. Збірник наукових праць. 2026;2(84):129-133. doi:10.26906/sunz.2026.2.129
BibTeX
@article{Lychkatyi2026,
author = {Lychkatyi, Oleksandr and Povoroznyuk, Anatoliy},
title = {МУЛЬТИФРАКТАЛЬНИЙ АНАЛІЗ МАМОГРАФІЧНИХ ЗОБРАЖЕНЬ},
journal = {Системи управління, навігації та зв’язку. Збірник наукових праць},
year = {2026},
volume = {2},
number = {84},
pages = {129--133},
doi = {10.26906/sunz.2026.2.129}
}
EUVIMED