POPOOLA, S. Boundary effects in signal processing: From classical problems to new opportunities [online]. Brno: Vysoké učení technické v Brně. Fakulta strojního inženýrství. 2023.

Posudek vedoucího

Cicone, Antonio

In this work, James focuses on the boundary problems of signal processing methods. He reviews classical approaches for windowing data and then he presents recent advancements in the pre-extension of signals to reduce boundary effects.

Posudek oponenta

Druckmüllerová, Hana

The master thesis deals with signals processing. It starts with the overview of the classical methods in Chapter 1, followed by the introduction to boundary effect problems and the classical approaches in Chapter 2, methods developed in the last decades for non-stationary signals in Chapter 3 and finally the most recent methods of boundary effects elimination in Chapter 4. I appreciate the general introduction to signal processing and analysis and to the boundary effect problem, which is suitable even for a mathematician that is not familiar with signals. What I miss here the most, is a more rigorous mathematical description with more formulae and at least some definitions and theorems, which would also improve consistency of the notation (as an example T is the fundamental time period in the introductory part of Chapter 1, while in formula (1.7.30) on page 34 it is the contribution of frequency f at time t, see also formula (4.4.7) on page 63 - j is the imaginary unit in Chapters 1 and 2, later in Chapters 3 and 4 used for indexing, here i seems to be both an index and the imaginary unit). A review of the modern methods to eliminate the boundary effect problem is the contribution of the author. Even though the thesis was meant as a review, I would appreciate it if the author prepared more pictures on his own. Almost all pictures were taken from cited sources, which resulted in inconsistent fonts and sometimes notations (e.g. figure 3.3 on page 51). It is clear that the author did a huge amount of work working with many sources, however, the quality of citations is poor. Internet sources lack links and dates of citation, some citations even seem to be unsearchable ([120, 121] for instance), some have serious mistakes (the author of [135] is P. Welch, not Welch and Peter, [105] has serious mistakes in authors), the citation style is inconsistent ([126] is a book cited as an incomplete web link). Some sources are even not cited at all, e.g. Bergland 1969, Strang 1993 from page 28. After balancing the author’s contribution and the overall quality of the thesis, I recommend it to be admitted to the defense.

eVSKP id 149983