Přístupnostní navigace
E-application
Search Search Close
Bachelor's Thesis
Author of thesis: Roman Ivakhniuk
Acad. year: 2025/2026
Supervisor: Mgr. Jan Faltýnek, Ph.D.
Reviewer: Ing. Jan Turčínek, Ph.D.
This bachelor's thesis presents a software solution for transcribing 17th-century historical music printed by Pierre Ballard. Currently, evaluating the sound of these digitized compositions is impossible before investing significant time into manual transcription. To solve this, this work implements a multi-stage Optical Music Recognition (OMR) pipeline using YOLO neural networks. The software, named nanoScore, breaks the transcription process into four steps: detecting staves, detecting musical symbols, classifying their vertical positions, and algorithmically reconstructing the polyphonic score into a MusicXML format. To handle occasional neural network prediction errors, an interactive semiautomatic mode allows users to correct mistakes before the final score is generated. Evaluation of the trained models shows high accuracy. Staff and symbol detection reached an mAP50 of 0.995 and 0.927, respectively, and position classification achieved a Top-1 accuracy of 0.982. When tested on a 384-page partbook, the software achieved rhythmic synchronization for 71.5% of the measures across all four voices. Running efficiently on a standard consumer CPU, the pipeline provides musicologists and laypersons with an accessible tool to generate an approximate sound assessment from images, allowing them to hear and evaluate a piece before committing to a manual transcription.
Historical music notation, Optical Music Recognition (OMR), object detection, image classification, YOLO (You Only Look Once), score reconstruction, MusicXML
Date of defence
16.06.2026
Result of the defence
Defended (thesis was successfully defended)
Grading
A
Process of defence
Student seznámil komisi s výsledky své bakalářské práce a odpověděl na otázky oponenta. Následovala diskuze k závěrečné práci a otázky komise: Rozpoznání různých notových klíčů a jejich transpozice. Kontrola délky jednotlivých taktů. Student odpověděl na všechny doplňující dotazy ke spokojenosti komise.
Language of thesis
English
Faculty
Fakulta strojního inženýrství
Department
Institute of Automation and Computer Science
Study programme
Engineering (B-STR-P)
Specialization
Applied Computer Science and Control (AIŘ)
Composition of Committee
doc. Ing. Lukáš Novotný, Ph.D. (předseda) doc. Ing. Miloš Hammer, CSc. (místopředseda) Ing. et Ing. Stanislav Lang, Ph.D. (člen) Mgr. Jan Faltýnek, Ph.D. (člen) Ing. Petr Lošák, Ph.D. (člen) Ing. Filip Kšica, Ph.D. (člen) doc. Ing. Ivan Švarc, CSc. (člen) doc. Ing. Radek Vlach, Ph.D. (člen) Ing. Luděk Janák (člen)
Supervisor’s reportMgr. Jan Faltýnek, Ph.D.
Grade proposed by supervisor: A
Reviewer’s reportIng. Jan Turčínek, Ph.D.
Grade proposed by reviewer: A
Responsibility: Mgr. et Mgr. Hana Odstrčilová