EngravingGNN: A Hybrid Graph Neural Network for End-to-End Piano Score Engraving

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Karystinaios, Emmanouil, Foscarin, Francesco, Widmer, Gerhard
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866914052860542976
author Karystinaios, Emmanouil
Foscarin, Francesco
Widmer, Gerhard
author_facet Karystinaios, Emmanouil
Foscarin, Francesco
Widmer, Gerhard
contents This paper focuses on automatic music engraving, i.e., the creation of a humanly-readable musical score from musical content. This step is fundamental for all applications that include a human player, but it remains a mostly unexplored topic in symbolic music processing. In this work, we formalize the problem as a collection of interdependent subtasks, and propose a unified graph neural network (GNN) framework that targets the case of piano music and quantized symbolic input. Our method employs a multi-task GNN to jointly predict voice connections, staff assignments, pitch spelling, key signature, stem direction, octave shifts, and clef signs. A dedicated postprocessing pipeline generates print-ready MusicXML/MEI outputs. Comprehensive evaluation on two diverse piano corpora (J-Pop and DCML Romantic) demonstrates that our unified model achieves good accuracy across all subtasks, compared to existing systems that only specialize in specific subtasks. These results indicate that a shared GNN encoder with lightweight task-specific decoders in a multi-task setting offers a scalable and effective solution for automatic music engraving.
format Preprint
id arxiv_https___arxiv_org_abs_2509_19412
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle EngravingGNN: A Hybrid Graph Neural Network for End-to-End Piano Score Engraving
Karystinaios, Emmanouil
Foscarin, Francesco
Widmer, Gerhard
Graphics
Artificial Intelligence
This paper focuses on automatic music engraving, i.e., the creation of a humanly-readable musical score from musical content. This step is fundamental for all applications that include a human player, but it remains a mostly unexplored topic in symbolic music processing. In this work, we formalize the problem as a collection of interdependent subtasks, and propose a unified graph neural network (GNN) framework that targets the case of piano music and quantized symbolic input. Our method employs a multi-task GNN to jointly predict voice connections, staff assignments, pitch spelling, key signature, stem direction, octave shifts, and clef signs. A dedicated postprocessing pipeline generates print-ready MusicXML/MEI outputs. Comprehensive evaluation on two diverse piano corpora (J-Pop and DCML Romantic) demonstrates that our unified model achieves good accuracy across all subtasks, compared to existing systems that only specialize in specific subtasks. These results indicate that a shared GNN encoder with lightweight task-specific decoders in a multi-task setting offers a scalable and effective solution for automatic music engraving.
title EngravingGNN: A Hybrid Graph Neural Network for End-to-End Piano Score Engraving
topic Graphics
Artificial Intelligence
url https://arxiv.org/abs/2509.19412