Modelovanie a kvantitatívna analýza dynamiky mechanických vlastností umelého cievneho systému

Abstract

The aim of this diploma thesis is analysis, modeling and quantitative evaluation of the dynamics of the artificial blood vessel system. The analysis contains the use of algorithms to detect a vessel from a video. The thesis implements segmentation algorithms with using methods of region growing, fuzzy soft thresholding and Otsu thresholding. The segmentation are objectively compared with the gold standard by comparative analysis. The results of this analysis were the basis for creation of optimal segmentation model of the vessel system from the video. Within the thesis were extracted geometric parameters of the blood vessel (area, perimeter, corrugation, thickness) describing its mechanical properties. Subsequently, correlation coefficients were calculated expressing degree of dependence between features from the video and course of the pressure curve from external sensor. The thesis also includes a graphical user interface to demonstrate the segmentation and features extraction from video.