Publications Vojtech Dvorak

Publications

Fast Optical Music Recognition Using the YOLO Platform

This bachelor thesis emphasizes practical, scalable OMR systems capable of processing large music archives rather than purely theoretical models. It introduces a modular OMR pipeline based on YOLO11 object detection, separating detection and interpretation to improve efficiency, accuracy, and tunability. Experiments on the OmniOMR and OLiMPiC datasets show notable gains in speed and detection performance, with simplified MusicXML outputs and open-source tools released for future work.

DVOŘÁK, Vojtěch. Rychlé rozpoznávání notopisů pomocí platformy YOLO. Bakalářská práce, vedoucí Mayer, Jiří. Praha: Univerzita Karlova, Matematicko-fyzikální fakulta, Ústav formální a aplikované lingvistiky, 2025. https://dspace.cuni.cz/handle/20.500.11956/200867

Staff Layout Analysis Using the YOLO Platform

In this paper, we bring an update on the performance of OMR layout analysis with the state-of-the-art YOLO platform. Compared to the MeasureDetector (the main publicly available tool for layout analysis), it achieves a similar or better accuracy across both in-domain and out-of-domain tests over three different datasets that we harmonized, it is more than 20x faster, and requires more than 4 times less memory.

Vojtěch Dvořák, Jan jr. Hajič, and Jiří Mayer. Staff Layout Analysis Using the YOLO Platform. In Jorge Calvo-Zaragoza, Alexander Pacha, and Elona Shatri, editors, Proceedings of the 6th International Workshop on Reading Music Systems, pages 18-22, Online, 2024. https://arxiv.org/abs/2411.15741