Research map / Image generation / Autoregressive and latent generation
End-to-End Full-Page Optical Music Recognition for Pianoform Sheet Music
Antonio Ríos-Vila, Jorge Calvo-Zaragoza, David Rizo, Thierry Paquet
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| replace CNN backbone with ConvNeXt | SMT (bekern) · FP-GrandStaff test (synthetic) | SER: 19.7 -> 6.9 | Improved |
| use ekern encoding instead of raw kern | SMT-ConvNeXt · FP-GrandStaff test (synthetic) | SER: 9.7 -> 8.5 | Improved |
| use bekern encoding instead of raw kern | SMT-ConvNeXt · FP-GrandStaff test (synthetic) | SER: 9.7 -> 6.9 | Improved |
| use ekern encoding instead of raw kern | SMT-CNN · FP-GrandStaff test (synthetic) | CER: 15.8 -> 17.3 | Hurt |
| add system-level pre-training (ablation) | SMT-ConvNeXt · FP-GrandStaff test (synthetic) | SER: 12.4 -> 6.9 | Improved |
| add incremental curriculum learning (ablation) | SMT-ConvNeXt · FP-GrandStaff test (synthetic) | SER: 9.8 -> 6.9 | Improved |
| add both pre-training and curriculum to baseline (ablation) | SMT-ConvNeXt · FP-GrandStaff test (synthetic) | SER: 42.6 -> 6.9 | Improved |
| add system-level pre-training only (ablation) | SMT-ConvNeXt · FP-GrandStaff test (synthetic) | SER: 42.6 -> 9.8 | Improved |
| add 1D positional encoding vs none (ablation) | SMT-ConvNeXt · FP-GrandStaff test (synthetic) | SER: 78.4 -> 15.57 | Improved |
| use 2D instead of 1D positional encoding (ablation) | SMT-ConvNeXt · FP-GrandStaff test (synthetic) | SER: 15.57 -> 6.9 | Improved |
| train on more systems per page (max 1 -> 5) and test on 2 systems (ablation) | SMT-ConvNeXt · Synthetic curriculum matrix | SER: 48.7 -> 4.91 | Improved |
| fine-tune on real scores (ablation) | SMT-ConvNeXt · Mozarteum (real) | SER: 38.9 -> 14.1 | Improved |
Closest papers by meaning
- Transcoda: End-to-End Zero-Shot Optical Music Recognition via Data-Centric Synthetic Training (NeurIPS 2026 · similarity 0.93)
- Toward Multimodal Sheet Music Recognition and Understanding (NeurIPS 2026 · similarity 0.91)
- LEGATO: Large-scale End-to-end Generalizable Approach to Typeset OMR (ICLR 2026 · similarity 0.91)
- Aria-MIDI: A Dataset of Piano MIDI Files for Symbolic Music Modeling (ICLR 2025 · similarity 0.85)
- Rubato: Transcribing Piano Music with Timestamps (NeurIPS 2026 · similarity 0.84)
- N-gram Unsupervised Compoundation and Feature Injection for Better Symbolic Music Understanding (AAAI 2024 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
