Saved in:
Bibliographic Details
Main Authors: Dratschuk, Daniel, Swoboda, Paul
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2605.10835
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917481052897280
author Dratschuk, Daniel
Swoboda, Paul
author_facet Dratschuk, Daniel
Swoboda, Paul
contents Optical Music Recognition (OMR), the task of transcribing sheet music into a structured textual representation, is currently bottlenecked by a lack of large-scale, annotated datasets of real scans. This forces models to rely on either few-shot transfer or synthetic training pipelines that remain overly simplistic. A secondary challenge is encoding non-uniqueness: in the popular Humdrum **kern format for transcribing music, multiple different text encodings can render into the same visual sheet music. This one-to-many mapping creates a harder learning task and introduces high uncertainty during decoding. We propose Transcoda, an OMR system built on (i) an advanced synthetic data generation pipeline, (ii) a normalization of the **kern encoding to enforce a unique normal form and (iii) grammar-based decoding to ensure the syntactic correctness of the output. This approach allows us to train a compact 59M-parameter model in just 6 hours on a single GPU that outperforms billion-parameter baselines. Transcoda achieves the best score among state of the art baselines on a newly curated benchmark of synthetically rendered scores at 18.46% OMR-NED (compared to 43.91% for the next-best system, Legato) and reduces the error rate on historical Polish scans to 63.97% OMR-NED (down from 80.16% for SMT++).
format Preprint
id arxiv_https___arxiv_org_abs_2605_10835
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Transcoda: End-to-End Zero-Shot Optical Music Recognition via Data-Centric Synthetic Training
Dratschuk, Daniel
Swoboda, Paul
Computer Vision and Pattern Recognition
Machine Learning
Optical Music Recognition (OMR), the task of transcribing sheet music into a structured textual representation, is currently bottlenecked by a lack of large-scale, annotated datasets of real scans. This forces models to rely on either few-shot transfer or synthetic training pipelines that remain overly simplistic. A secondary challenge is encoding non-uniqueness: in the popular Humdrum **kern format for transcribing music, multiple different text encodings can render into the same visual sheet music. This one-to-many mapping creates a harder learning task and introduces high uncertainty during decoding. We propose Transcoda, an OMR system built on (i) an advanced synthetic data generation pipeline, (ii) a normalization of the **kern encoding to enforce a unique normal form and (iii) grammar-based decoding to ensure the syntactic correctness of the output. This approach allows us to train a compact 59M-parameter model in just 6 hours on a single GPU that outperforms billion-parameter baselines. Transcoda achieves the best score among state of the art baselines on a newly curated benchmark of synthetically rendered scores at 18.46% OMR-NED (compared to 43.91% for the next-best system, Legato) and reduces the error rate on historical Polish scans to 63.97% OMR-NED (down from 80.16% for SMT++).
title Transcoda: End-to-End Zero-Shot Optical Music Recognition via Data-Centric Synthetic Training
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2605.10835