Návrh a realizace softwarové aplikace pro analýzu neurologických abnormalit z obrazů magnetické rezonance

Abstract

This work focuses on the analysis of images from magnetic resonance imaging (MRI) with the aim of creating supportive images for diagnosing patients with neurological abnormalities, specifically focal cortical dysplasia (FCD). The initial chapters are dedicated to a literature review on the anatomy, physiology, and pathophysiology of the brain, as well as the principles of MRI and methods for analyzing MRI images. This theoretical part provides readers with sufficient background to understand the subsequent practical section. The practical section begins with a description of the datasets used and the preprocessing of MRI images, which are essential for further analysis steps leading to the creation of supportive diagnostic images (convolutional images and junction images). These supportive images yield good results for patients with FCD, highlighting pathologies in both patients with FCD type 2 from a public database and patients with FCD type 1 from University Hospital of Ostrava. Average cortical thicknesses of affected lobes are also calculated to quantify the results. The results are discussed at the end, including the limitations of the work.