CHUDÁREK, A. Analyzing a person’s handwriting for recognizing his/her emotional state [online]. Brno: Vysoké učení technické v Brně. Fakulta informačních technologií. 2024.
The student has completed all the objectives specified in the project description. He has developed a machine learning-based approach to analyze handwriting for the detection of stress, anxiety and depression. He has tested it with two different datasets and the results are promising. The student has provided detailed results. It’s a very good work and thesis is easy to read and well written. It would have been better to also compare the results with methods mentioned in table 2.6.
| Kritérium | Známka | Body | Slovní hodnocení |
|---|---|---|---|
| Informace k zadání | The thesis was related to detection of stress, anxiety, and depression from the handwriting of a person. The work was challenging because it not only required knowledge in signal processing, image processing and machine learning but also required understanding of the mental health problems of stress, anxiety, and depression. The complexity of the project increased due to the multi-level classification, that is stress, anxiety, and depression. The student developed a machine learning model to achieve this objective and tested it on two different databases. The results are promising. | ||
| Práce s literaturou | The student has provided detailed literature review in chapter 2. The chapter starts with description of emotions and the corresponding handwriting analysis. Then the various datasets are discussed that have been used by various researchers in this domain. After discussing the datasets, the pre-processing and feature extraction is discussed followed by the classification. Finally, many articles are presented which have attempted to provide a solution to this problem. Table 2.6 summarizes and compares such articles. Overall, it’s a well written chapter. | ||
| Aktivita během řešení, konzultace, komunikace | The student had regular weekly meetings throughout the two semesters while working on the thesis. He came to meetings on time and I found him to be a punctual and hardworking student. We had two collaborators on this project and the student satisfactorily interacted with them too. | ||
| Aktivita při dokončování | The student regularly consulted me before the submission of the thesis. The work was challenging because two datasets were involved as well as collaboration with two external partners were involved. The student was able to complete all the development in time and was able to submit the thesis within the given timeframe. | ||
| Publikační činnost, ocenění | The thesis is well written and contains content which is publishable. The student is willing to assist in preparing the draft for a journal paper. The journal paper is being prepared and will be submitted to a Q1 journal. |
While on one hand the thesis does not fulfill the entire assignment, on the other hand it presents significantly more advanced implementation and correspondingly more comprehensive evaluation, which are introduced in a well structured and easy to understand technical report written in English. The presented results are of a high quality not only compared to the common master thesis level but also compared to existing scientific literature. Therefore, I propose grade A / excellent.
| Kritérium | Známka | Body | Slovní hodnocení |
|---|---|---|---|
| Náročnost zadání | I consider the assignment to be relatively difficult due to its interdisciplinary nature and a limited number of related works. Moreover, as documented by the student in Sections 2.5 and 2.6, the use and/or evaluation of machine learning approaches in existing literature is often simplified. Therefore, the student had to pay special attention to correct application and evaluation of machine learning techniques in the context of the addressed problem. | ||
| Rozsah splnění požadavků zadání | Unfortunately, the reviewed thesis does not contain the evaluation of "computational performance and memory usage," which is required - among other evaluations - in item 6 of the assignment. On the other hand, the set of implemented Python scripts rather represents a whole framework for training machine learning models than a single algorithm for the interpretation of emotion and mood from a person’s handwriting, which is required in item 4 of the assignment. In alignment with that, the presented evaluation campaign significantly exceeds the requirements set by the assignment. Thus, it seems to me that the missing evaluation was omitted primarily because of focusing on other aspects of the thesis. Moreover, the student should be able to present the missing evaluation at the thesis defence, which I am actually asking for (see Questions for Defence). | ||
| Rozsah technické zprávy | The length of the thesis is within the usual range. The details of unsuccessful experiments and a detailed evaluation of utilized features were correctly included as appendices. | ||
| Prezentační úroveň technické zprávy | 90 | In general, the whole thesis is well structured and easy to understand. This is especially true for Chapters 1-3 and 7, the majority of which were subject to the semestral defence at the end of the winter semester. In my opinion, the key components facilitating the understandability of these parts of the thesis are summary Sections 2.6 and 3.6. Throughout the thesis, there are some presentation-related issues that could be addressed (a few misleading references, concurrent use of terms "recall" and "sensitivity" for the same metric, difficult to follow numbering of experiments, missing references to and descriptions of Tables 6.1, 6.3, and 6.4, a duplicate description of Experiment 4 in Section A.1.1). Nevertheless, considering the length of the thesis, these issues represent just a tiny portion of its content. | |
| Formální úprava technické zprávy | 95 | The thesis is of very high quality with respect to both typography and language. Regarding typography, the only little issues I encountered are inconsistent capitalization of Section titles and the use of a standard space instead of a non-breaking space in front of references to Sections and Bibliography. The most common issues regarding language are typos and (typically missing) commas. However, the use of language in this thesis written in English is highly above the average of other theses I read, even considering theses written in Czech. Actually, I identified proportionally more language issues in the Extended Abstract, which is written in Czech, than in the rest of the thesis. | |
| Práce s literaturou | 100 | The Bibliography section contains references to relevant journal articles, conference papers, as well as other online sources. These references are complete and they follow common citation standards. Apart from various publication resources, the student also utilized publicly available datasets and an existing library for the extraction of a variety of common handwriting features. All the utilized resources are correctly introduced in the thesis. | |
| Realizační výstup | 98 | The set of implemented Python scripts is sufficiently commented and fully working. Project's repository also contains a basic documentation, which lists required dependencies and contains instructions regarding installation and usage. | |
| Využitelnost výsledků | The reviewed thesis introduces a novel - more comprehensive and more systematically evaluated - machine learning approach to the interpretation of emotion and mood from a person’s handwriting. It therefore significantly extends the previously published results in this area and also presents some original results and perspectives related to this field. In my opinion, the core of the thesis is worth publishing in a relevant journal or at a relevant conference. |
eVSKP id 153394