Identifikace a kvantifikace kolagenních vláken z mikroskopických obrazů

Abstract

This bachelor thesis is focused on identification and quantification collagen fibers from microscopic images. The aim of this thesis was detecting collagen fibers and specify quantification methods. In the theoretical part, there are outlined the basic tenets of anatomy of tissues and blood vessels and there is devotion to the theory of image segmentation, which is the key to this bachelor thesis. This thesis also includes a recherche methods for quantification waviness of collagen fibers. The practical part is focused on a preprocessing of microscopic image, detection of collagen fibers in images and then determination quantification method – waviness and thickness. The method of segmentation is used to detect fibers in images. The active contours method was depicted by image segmentation. The processes that preceded it during processing were presented as well in this bachelor thesis, such as filtering using a median filter, not only the image segmentation. Subsequently, the results are compared with gold standards of collagen fibers, which were manually marked. The designed procedures were tested on a dataset of thirteen microscopic images.