# From Image to Music Language: A Two-Stage Structure Decoding Approach for Complex Polyphonic OMR

arXiv: https://arxiv.org/abs/2604.20522 · HTML: https://mlsanity.com/paper/2604.20522/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 8 of its 31 references that are in the map.
- First posted on arXiv 22 Apr 2026.

## Builds on (most cited first, 8 of 8)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Music Transformer](https://mlsanity.com/paper/1809.04281/) (2018)
- [Grammar-Constrained Decoding for Structured NLP Tasks without Finetuning](https://mlsanity.com/paper/2305.13971/) (2023)
- [Understanding Optical Music Recognition](https://mlsanity.com/paper/1908.03608/) (2019)
- [DeepScores -- A Dataset for Segmentation, Detection and Classification of Tiny Objects](https://mlsanity.com/paper/1804.00525/) (2018)
- [A Unified Representation Framework for the Evaluation of Optical Music Recognition Systems](https://mlsanity.com/paper/2312.12908/) (2023)
- [Toward a More Complete OMR Solution](https://mlsanity.com/paper/2409.00316/) (2024)
- [Residual Recurrent CRNN for End-to-End Optical Music Recognition on Monophonic Scores](https://mlsanity.com/paper/2010.13418/) (2020)

## Abstract (arXiv)

We propose a new approach for a practical two-stage Optical Music Recognition (OMR) pipeline, with a particular focus on its second stage. Given symbol and event candidates from the visual pipeline, we decode them into an editable, verifiable, and exportable score structure. We focus on complex polyphonic staff notation, especially piano scores, where voice separation and intra-measure timing are the main bottlenecks. Our approach formulates second-stage decoding as a structure decoding problem and uses topology recognition with probability-guided search (BeadSolver) as its core method. We also describe a data strategy that combines procedural generation with recognition-feedback annotations. The result is a practical decoding component for real OMR systems and a path to accumulate structured score data for future end-to-end, multimodal, and RL-style methods.