Title of the Thesis:
"Computer Vision and Machine Learning Analysis of
Atomic-Scale Microscopy Images”
Defense committee:
Luca Ghiringhelli, Friedrich-Alexander Universität Erlangen, DE (reviewer)
Stefan Förster, Martin-Luther-Universität Halle-Wittenberg, DE (reviewer)
Cesare Franchini (supervisor)
Thomas Pichler (chair)
Abstract:
The development of new computational methods often arises from the need to accomplish specific tasks that were previously unfeasible or required a strenuous amount of manual effort. In the field of materials science, in particular microscopy, many steps in the experimental workflow are still manually or semi-manually executed, heavily relying on the user’s expertise. This doctoral project focused on the development and application of workflows for the analysis and processing of microscopy images, focusing on crystalline, quasicrystalline pattern recognition, and image denoising. Each method has been tested across different case scenarios, involving various microscopy techniques and a wide range of noise levels and image artifacts. The limitations of each methodology have been carefully identified and discussed.
