Fractal analysis of human cerebellum based on magnetic resonance imaging data: pixel dilating method
Year: 2020
Type: article
Study design: Observational Study
Abstract: Background. The use of fractal analysis as a morphometric method allows to investigate biological structures that have the properties of fractals, including the human cerebellum. Adaptation of fractal analysis techniques to assess the state of brain structures using magnetic resonance imaging is an important area of modern morphology. Objective: to determine the values of the fractal dimension (FD) of the cerebellar tissue measured by magnetic resonance imaging using the pixel dilation. Methods. The study was conducted on magnetic resonance images of the brain of 120 relatively healthy patients. Fractal analysis of digital magnetic resonance images of the cerebellum was performed using the method of pixel dilation in the author's modification. Results. It was found that the average value of FD of the cerebellar vermis on the midsagittal section on T2 weighed images with a brightness threshold of 100 was 1.691 ± 0.01. The FD values of the cerebellar hemisphere tissue were: in the paravermal zone on the left 1.683 ± 0.01, on the right 1.685 ± 0.01; in the central zone of the hemisphere on the left 1.679 ± 0.01, on the right 1.672 ± 0.01; on the lateral zone of the hemisphere on the left 1.665 ± 0.01, on the right 1.682 ± 0.01. These values do not differ statistically significantly in the symmetrical areas of the right and left hemispheres and do not differ from the FD values of the cerebellar vermis. Conclusion. The method of pixel dilation of fractal analysis can be used for morphometric study of magnetic resonance images of the brain. Fractal analysis allows an objective assessment of the morphofunctional condition of the cerebellum, which can be used to diagnose various diseases of the cerebellum and other CNS structures.
Source: Morphologia
Authors: N. I. Maryenko, O. Yu. Stepanenko
Institution: Kharkiv National Medical University
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W3116151663