Optical Music Recognition of Jazz Lead Sheets

Juan Carlos Martinez-Sevilla; Francesco Foscarin; Patricia Garcia-Iasci; David Rizo; Jorge Calvo-Zaragoza; Gerhard Widmer

doi:10.48550/arXiv.2509.05329

Optical Music Recognition of Jazz Lead Sheets

Juan Carlos Martinez-Sevilla^*
, Francesco Foscarin
, Patricia Garcia-Iasci
, David Rizo
, Jorge Calvo-Zaragoza
, Gerhard Widmer

^*Corresponding author for this work

Institute of Computational Perception

Research output: Working paper and reports › Preprint

Abstract

In this paper, we address the challenge of Optical Music Recognition (OMR) for handwritten jazz lead sheets, a widely used musical score type that encodes melody and chords. The task is challenging due to the presence of chords, a score component not handled by existing OMR systems, and the high variability and quality issues associated with handwritten images. Our contribution is two-fold. We present a novel dataset consisting of 293 handwritten jazz lead sheets of 163 unique pieces, amounting to 2021 total staves aligned with Humdrum **kern and MusicXML ground truth scores. We also supply synthetic score images generated from the ground truth. The second contribution is the development of an OMR model for jazz lead sheets. We discuss specific tokenisation choices related to our kind of data, and the advantages of using synthetic scores and pretrained models. We publicly release all code, data, and models.

Original language	English
Number of pages	7
DOIs	https://doi.org/10.48550/arXiv.2509.05329
Publication status	Published - 31 Aug 2025

Publication series

Name	arXiv.org
No.	2509.05329

Fields of science

102003 Image processing
102001 Artificial intelligence
202002 Audiovisual media
102 Computer Sciences
102015 Information systems

JKU Focus areas

Digital Transformation

Access to Document

10.48550/arXiv.2509.05329

1 Conference proceedings

Optical Music Recognition of Jazz Lead Sheets
Martinez-Sevilla, J. C., Foscarin, F., Garcia-Iasci, P., Rizo, D., Calvo-Zaragoza, J. & Widmer, G., 2025, Proceedings of the 26th International Society for Music Information Retrieval Conference (ISMIR), 2025. 1 ed.
Research output: Chapter in Book/Report/Conference proceeding › Conference proceedings › peer-review

Cite this

@techreport{9d198bda75764af9833591d4846f2805,

title = "Optical Music Recognition of Jazz Lead Sheets",

abstract = "In this paper, we address the challenge of Optical Music Recognition (OMR) for handwritten jazz lead sheets, a widely used musical score type that encodes melody and chords. The task is challenging due to the presence of chords, a score component not handled by existing OMR systems, and the high variability and quality issues associated with handwritten images. Our contribution is two-fold. We present a novel dataset consisting of 293 handwritten jazz lead sheets of 163 unique pieces, amounting to 2021 total staves aligned with Humdrum **kern and MusicXML ground truth scores. We also supply synthetic score images generated from the ground truth. The second contribution is the development of an OMR model for jazz lead sheets. We discuss specific tokenisation choices related to our kind of data, and the advantages of using synthetic scores and pretrained models. We publicly release all code, data, and models.",

author = "Martinez-Sevilla, \{Juan Carlos\} and Francesco Foscarin and Patricia Garcia-Iasci and David Rizo and Jorge Calvo-Zaragoza and Gerhard Widmer",

year = "2025",

month = aug,

day = "31",

doi = "10.48550/arXiv.2509.05329",

language = "English",

series = "arXiv.org",

number = "2509.05329",

type = "WorkingPaper",

}

TY - UNPB

T1 - Optical Music Recognition of Jazz Lead Sheets

AU - Martinez-Sevilla, Juan Carlos

AU - Foscarin, Francesco

AU - Garcia-Iasci, Patricia

AU - Rizo, David

AU - Calvo-Zaragoza, Jorge

AU - Widmer, Gerhard

PY - 2025/8/31

Y1 - 2025/8/31

N2 - In this paper, we address the challenge of Optical Music Recognition (OMR) for handwritten jazz lead sheets, a widely used musical score type that encodes melody and chords. The task is challenging due to the presence of chords, a score component not handled by existing OMR systems, and the high variability and quality issues associated with handwritten images. Our contribution is two-fold. We present a novel dataset consisting of 293 handwritten jazz lead sheets of 163 unique pieces, amounting to 2021 total staves aligned with Humdrum **kern and MusicXML ground truth scores. We also supply synthetic score images generated from the ground truth. The second contribution is the development of an OMR model for jazz lead sheets. We discuss specific tokenisation choices related to our kind of data, and the advantages of using synthetic scores and pretrained models. We publicly release all code, data, and models.

AB - In this paper, we address the challenge of Optical Music Recognition (OMR) for handwritten jazz lead sheets, a widely used musical score type that encodes melody and chords. The task is challenging due to the presence of chords, a score component not handled by existing OMR systems, and the high variability and quality issues associated with handwritten images. Our contribution is two-fold. We present a novel dataset consisting of 293 handwritten jazz lead sheets of 163 unique pieces, amounting to 2021 total staves aligned with Humdrum **kern and MusicXML ground truth scores. We also supply synthetic score images generated from the ground truth. The second contribution is the development of an OMR model for jazz lead sheets. We discuss specific tokenisation choices related to our kind of data, and the advantages of using synthetic scores and pretrained models. We publicly release all code, data, and models.

U2 - 10.48550/arXiv.2509.05329

DO - 10.48550/arXiv.2509.05329

M3 - Preprint

T3 - arXiv.org

BT - Optical Music Recognition of Jazz Lead Sheets

ER -