The aim of this thesis is to propose, implement, and evaluate
an automatic method for the segmentation of cell nucleus
instances in large three-dimensional fluorescence microscopy
image data. The proposed segmentation method uses a fully
convolutional neural network to predict a binary mask of cell
nuclei. Individual cell nucleus instances are then segmented
using the Watershed algorithm. The thesis contains a
description of the individual steps of the segmentation method,
a justification for the chosen approach, as well as a
quantitative evaluation of the results. An implementation of
the developed method in the Python programming language is a
part of the thesis.