X-Ray Image Analysis and Processing

Abstract

The aim of this master thesis was to design and test new machine learning methods designed to remove unwanted parts in X-ray images, in the case of this work it is mainly ribs. This modification should help with the subsequent analysis. The work focuses not only on the design of the new methods but also on their inter-comparison over both the annotated dataset and the dataset provided by Olomouc University Hospital. In addition to the results for each approach, I would like to mention the high performance of the proposed generative adversarial network, which achieves excellent results compared to other approaches.