Data-Driven Glyph Positioning for Handwritten Music Synthesis

A Graph Neural Network trained on the Music Notation Graph structure to predict context-dependent glyph offsets cut positioning RMSE by an additional 50 to 85% on stems and beams compared with empirical means, according to an arXiv paper (2610.28899v1) on data-driven glyph positioning for handwritten music synthesis. The authors extracted inter-glyph distributions from roughly 285,000 annotated relationships across MUSCIMA++ and OmniOMR, found several hard-coded Smashcima parameters deviate systematically from real handwriting, and showed that replacing those constants with empirical means reduced positioning RMSE by up to 60%. In downstream evaluation, a YOLOv8 detector pretrained on GNN-positioned synthetic data achieved +30% mAP50 over the baseline and +11% over empirical means when fine-tuned on 200 real handwritten staves, while cross-dataset analysis attributed spacing differences to note density (3.3) rather than style and found duration-proportional spacing does not hold in handwriting (R2 = 0.001).

arXiv:2610.28899v1 Announce Type: new Abstract: Synthesised data is a common strategy to address the lack of annotated handwritten music scores in Optical Music Recognition OMR . Smashcima is a recent synthesis framework that assembles pages of handwritten scores by sampling glyph images from MUSCIMA++ and positioning them with hard-coded rules. In this work we ask whether these rules are grounded empirically and whether a learned model for context-dependent positioning can further improve synthesis quality. We extract inter-glyph distributions from around 285 thousand annotated relationships across MUSCIMA++ and OmniOMR, finding that several key parameters deviate systematically from real handwriting. Replacing these constants with empirical means reduces positioning RMSE by up to 60%. We then train a Graph Neural Network GNN on the Music Notation Graph structure to predict context-dependent offsets, achieving an additional 50 to 85% RMSE reduction on stems and beams by modelling how neighbouring glyphs constrain placement. Cross-dataset analysis shows that spacing differences between the two corpora are explained by note density 3.3 , not style, and that duration-proportional spacing does not hold in handwriting R2 = 0.001 . In downstream evaluation, a YOLOv8 detector pretrained on GNN-positioned synthetic data achieves +30% mAP50 over the baseline and +11% over empirical means when fine-tuned on 200 real handwritten staves, showing that positioning accuracy matters for downstream OMR.