Přístupnostní navigace
E-application
Search Search Close
Doctoral Thesis
Author of thesis: Ing. Michal Gavenčiak
Acad. year: 2025/2026
Supervisor: prof. Ing. Jiří Mekyska, Ph.D.
Reviewers: prof. Ing. Peter Drotár, Ph.D., Dr. Ing. Jiří Přibil
Conventional assessment of graphomotor and handwriting difficulties is often constrained by subjectivity, high costs, and lengthy procedures, which can delay critical support for individuals in need. This dissertation aims to develop and validate a suite of novel computational methods for the objective, quantitative assessment of graphomotor, handwriting, and associated cognitive skills using on-line, digitally captured data. Four primary research avenues were pursued. First, a method based on the directional analysis of movement, enhanced by fractional-order calculus, was developed to assess graphomotor control in drawing tasks. Second, the principle of isochrony, the natural rhythm of handwriting, was quantified to create features assessing writing fluency. Third, a multimodal framework combining on-line handwriting with concurrent eye-tracking was investigated to assess cognitive-motor performance, yielding novel hand-eye coupling biomarkers. Finally, an LSTM-based autoencoder was explored as a data augmentation tool and preprocessing filter to enhance the feature space of limited clinical datasets. The results demonstrate the significant predictive gains of the proposed methods over conventional approaches. The directional analysis model achieved a balanced accuracy of 87\,\% in identifying children with graphomotor difficulties. The multimodal hand-eye tracking framework reached 90\,\% accuracy in assessing performance on a cognitive task. Isochrony-based features were shown to be highly predictive of handwriting proficiency, and the autoencoder preprocessing step provided an improvement over baseline model performance. Collectively, this research provides a validated toolkit of objective, powerful, and interpretable methods for analyzing the handwriting process. These contributions lay the groundwork for a new generation of decision support systems, promising more precise, efficient, and accessible assessment of handwriting and graphomotor skills.
On-line Handwriting Analysis, Graphomotor Skills, Developmental Dysgraphia, Machine Learning, Feature Extraction, Fractional Calculus, Multimodal Analysis, Eye-Tracking
Date of defence
17.12.2025
Result of the defence
Defended (thesis was successfully defended)
Process of defence
Disertant při prezentaci jasně a stručně vysvětlil dosažené výsledky vědecké práce. odpověděl na všechny otázky oponentů i členů komise. Obhajoba proběhla na velmi dobré úrovni.
Language of thesis
English
Faculty
Fakulta elektrotechniky a komunikačních technologií
Department
Department of Telecommunications
Study programme
Teleinformatics (DPC-TLI)
Composition of Committee
prof. Ing. Zdeněk Smékal, CSc. (předseda) doc. Ing. Otto Dostál, CSc. (člen) prof. Ing. Radim Burget, Ph.D. (člen) Ing. Jiří Přinosil, Ph.D. (člen) prof. Ing. Kamil Říha, Ph.D. (člen) prof. Ing. Peter Drotár, Ph.D. (člen) Dr. Ing. Jiří Přibil (člen)
Supervisor’s reportprof. Ing. Jiří Mekyska, Ph.D.
Reviewer’s reportprof. Ing. Peter Drotár, Ph.D.
Reviewer’s reportDr. Ing. Jiří Přibil
Responsibility: Mgr. et Mgr. Hana Odstrčilová