Metody segmentace obrazu založené na teorii grafů

Abstract

The aim of this thesis was to study selected graph-based image segmentation method. The text begins with general description of image analysis and it specifies the role of image segmentation in that process. Since there are many graph-based image segmentation methods already published couple of them were selected and briefly described their thoughts. The “Efficient graph-based image segmentation” method was chosen for deep learning. Method was explained in detail and algorithm was implemented in C++. In the last chapter are commented several results of image segmentation producing by the algorithm.