pith. sign in

arxiv: 2606.31811 · v1 · pith:66CYTZEZnew · submitted 2026-06-30 · 💻 cs.CV

MuSViT: A Foundation Vision Model for Sheet Music Representation

Carlos Penarrubia , Antonio Rios-Vila , Eliseo Fuentes-Martinez , Juan C. Martinez-Sevilla , Francisco J. Castellanos , Mar\'ia Alfaro-Contreras , Jorge Calvo-Zaragoza This is my paper

Pith reviewed 2026-07-01 06:10 UTC · model grok-4.3

classification 💻 cs.CV
keywords MuSViTsheet musicvision transformermasked autoencodersmusic score recognitionfoundation modelIMSLPsymbolic music
0
0 comments X

The pith

MuSViT produces vision embeddings that directly encode symbolic musical structure from sheet music pages.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper presents MuSViT as the first foundation vision model built specifically for sheet music by pre-training a Vision Transformer encoder with masked autoencoders on 9.7 million IMSLP pages. A two-stage curriculum begins with synthetic typeset scores before scaling to the full real-world corpus. In linear probing on music score recognition, symbol detection, and difficulty classification, MuSViT beats general-purpose vision encoders, while an embedding-transcription consistency test shows its representations align with music notation content where other models do not. The work positions MuSViT as a reusable backbone that transfers to multiple downstream sheet music tasks.

Core claim

MuSViT is a ViT encoder pre-trained via Masked Autoencoders on 9.7 million pages from the IMSLP using a two-stage curriculum of synthetic warm-up followed by large-scale training on the full corpus. Under linear probing it outperforms modern vision encoders on full-page and staff-level recognition, symbol detection, and difficulty classification; under fine-tuning it generally exceeds task-specific state-of-the-art methods. An embedding-transcription consistency analysis shows that MuSViT encodes symbolic musical structure directly in its representation space, unlike other encoders whose embeddings do not correlate with music notation content.

What carries the argument

Two-stage masked autoencoder pre-training curriculum on the IMSLP corpus that produces embeddings whose space correlates with symbolic musical notation content.

If this is right

  • MuSViT representations support strong performance on music score recognition and symbol detection even when the encoder remains frozen.
  • General-purpose vision encoders systematically miss the structured symbolic properties of musical notation.
  • Fine-tuning MuSViT yields gains over prior task-specific methods on the evaluated downstream tasks.
  • The model functions as a reusable foundation backbone for multiple sheet music understanding problems.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same pre-training pattern could be tested on other structured visual languages such as circuit diagrams or chemical structure drawings.
  • If the consistency analysis holds, MuSViT embeddings might support direct retrieval or alignment tasks between scores and audio without additional supervision.
  • Scaling the IMSLP pre-training further or adding multi-modal audio alignment could strengthen the symbolic encoding property.

Load-bearing premise

Pre-training via masked autoencoders on the IMSLP corpus with the two-stage curriculum produces representations whose embedding space correlates with symbolic musical content.

What would settle it

Run the embedding-transcription consistency analysis on MuSViT and several general vision encoders; if MuSViT embeddings show no higher correlation with transcribed notation content than the baselines, the central claim fails.

Figures

Figures reproduced from arXiv: 2606.31811 by Antonio Rios-Vila, Carlos Penarrubia, Eliseo Fuentes-Martinez, Francisco J. Castellanos, Jorge Calvo-Zaragoza, Juan C. Martinez-Sevilla, Mar\'ia Alfaro-Contreras.

Figure 1
Figure 1. Figure 1: Overview of MUSVIT. MUSVIT is pre-trained on diverse sheet music pages using Masked Autoencoders: patches are randomly masked and the model learns to reconstruct the missing content from the remaining visible context. We evaluate the generality of the learned representations by probing the encoder across four diverse downstream tasks: full-page and staff-level music score recognition, music symbol detectio… view at source ↗
Figure 2
Figure 2. Figure 2: MUSVIT performance across four downstream tasks. Left: Linear probing (frozen encoder)— MUSVIT (solid) consistently outperforms general-purpose vision encoders (dashed), demonstrating superior representation quality. Right: Fine-tuning—MUSVIT generally outperforms state-of-the-art methods (SoTA). Axes represent normalized performance on each task (higher is better); see Section 3 for detailed results. 1 In… view at source ↗
Figure 3
Figure 3. Figure 3: MUSVIT reconstruction example. Left: Masked music score image. Middle: MUSVIT reconstruction. Right: The original sheet music. The coloured rectangles highlight different recon￾struction components: Red indicates pitch sequence reconstruction (staff position); Blue represents clef reconstruction; Green marks rest reconstruction; and Purple denotes musical note sequence reconstruction (musical symbol aligne… view at source ↗
Figure 4
Figure 4. Figure 4: Representative pages from the IMSLP pre-training corpus, illustrating its visual diversity. The collection spans historical periods, notation systems, engraving styles, and musical textures. monophonic vocal lines to dense orchestral scores, across both modern typeset and historical engraving conventions—providing the visual diversity necessary for learning generalizable representations. Specif￾ically, we … view at source ↗
Figure 5
Figure 5. Figure 5: MUSVIT reconstruction examples. Each example row shows the masked input with 70% of patches removed (left), the MUSVIT reconstruction (middle), and the original sheet music region (right). A.3 Pre-Training Details Pre-training follows a two-stage curriculum. Stage 1 uses synthetic crops from DeepScoresV2 [40] at 512 × 512 resolution (1,024 patches per image, masking ratio 50%), serving as a structured warm… view at source ↗
Figure 6
Figure 6. Figure 6: Comparison between single-stage IMSLP training and the proposed two-stage curriculum. [PITH_FULL_IMAGE:figures/full_fig_p020_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Representative pages from the IMSLP pre-training corpus. The collection spans historical periods, notation systems (mensural, CWMN), engraving styles (handwritten, typeset), and musical textures (monophonic, polyphonic). A.7 Downstream Task Datasets We provide representative examples from each dataset used in the downstream evaluation, organized by task: • [PITH_FULL_IMAGE:figures/full_fig_p021_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Representative pages from the two full-page recognition datasets. [PITH_FULL_IMAGE:figures/full_fig_p022_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Representative staff images from the five staff-level recognition corpora. [PITH_FULL_IMAGE:figures/full_fig_p022_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Representative annotated score pages from [PITH_FULL_IMAGE:figures/full_fig_p023_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Representative score pages from the three score difficulty classification datasets. [PITH_FULL_IMAGE:figures/full_fig_p024_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: K-curves showing average embedding distance [PITH_FULL_IMAGE:figures/full_fig_p030_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: PCA activation maps for MUSVIT. Each pair shows a sheet music page (left) alongside the first-principal-component projection of patch embeddings from the final Transformer layer, rendered as a spatial heat map (right). Warm (red/orange) tones indicate high activation; cool (blue) tones indicate low activation. Twelve examples are shown across three panels, covering modern typeset scores, historical engrav… view at source ↗
Figure 14
Figure 14. Figure 14: PCA visualization of patch embeddings for four score pages. Each row shows, from left [PITH_FULL_IMAGE:figures/full_fig_p032_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: Additional PCA visualizations following the same layout as Fig. 14. The pattern is consistent [PITH_FULL_IMAGE:figures/full_fig_p033_15.png] view at source ↗
read the original abstract

Foundation models have transformed vision and language processing by providing rich, reusable representations that transfer across diverse tasks. Sheet music, as a visual encoding of musical language, lacks such a strong domain-specific backbone. We introduce MuSViT (Music Score Vision Transformer): the first foundation vision model for sheet music representation -- a ViT encoder pre-trained via Masked Autoencoders on 9.7 million pages from the IMSLP. To handle the complexity of real-world scores, we adopt a two-stage curriculum: a synthetic warm-up on typeset scores followed by large-scale training on the full IMSLP corpus. We evaluate MuSViT on four downstream tasks -- full-page and staff-level music score recognition, music symbol detection, and score difficulty classification -- under two scenarios: linear probing (frozen encoder) and fine-tuning. Under linear probing, MuSViT consistently outperforms modern vision encoders, revealing that general-purpose representations, regardless of scale, fall systematically short on the structured symbolic properties of musical notation. Under fine-tuning, MuSViT generally improves upon task-specific state-of-the-art methods. An additional embedding-transcription consistency analysis reveals that MuSViT encodes symbolic musical structure directly in its representation space -- unlike other encoders, whose embeddings do not correlate with music notation content. These results establish MuSViT as a foundation backbone for sheet music understanding.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 1 minor

Summary. The manuscript introduces MuSViT, the first foundation vision model for sheet music: a ViT encoder pre-trained via Masked Autoencoders on 9.7 million IMSLP pages using a two-stage curriculum (synthetic typeset warm-up followed by full real-world corpus). It evaluates the model on four downstream tasks (full-page and staff-level music score recognition, music symbol detection, score difficulty classification) under linear probing (frozen encoder) and fine-tuning, claiming consistent outperformance over modern general-purpose vision encoders in the linear-probing regime and general improvement over task-specific SOTA under fine-tuning. An embedding-transcription consistency analysis is presented to support the claim that MuSViT representations encode symbolic musical structure directly, unlike other encoders whose embeddings do not correlate with music notation content.

Significance. If the central empirical claims hold after methodological clarification, the work would constitute a meaningful contribution by establishing the first large-scale domain-specific vision foundation model for sheet music. The scale of the IMSLP pre-training corpus, the two-stage curriculum, and the dual linear-probing/fine-tuning evaluation protocol are positive elements that could provide a reusable backbone for music score understanding tasks. The linear-probing results, if robust, would usefully demonstrate that general-purpose encoders fall short on structured symbolic notation properties.

major comments (1)
  1. [Embedding-transcription consistency analysis] Embedding-transcription consistency analysis (abstract and corresponding results section): The strongest claim—that MuSViT encodes symbolic musical structure directly while other encoders do not—depends entirely on this analysis. No description is supplied of the alignment procedure between embeddings and transcribed content, the correlation or distance metric employed, whether transcription is performed by an independent symbol recognizer, or any controls that would isolate symbolic properties (pitch, rhythm, voice leading) from generic visual layout or glyph statistics. This detail is required to substantiate the claim.
minor comments (1)
  1. The abstract states 'consistent outperformance' and 'generally improves' without referencing specific tables, metrics, or statistical tests; the main text should include these with error bars or significance tests to allow verification of the reported trends.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for highlighting the need for methodological clarity in the embedding-transcription consistency analysis. We agree this section requires expansion to fully support the claims and will revise accordingly.

read point-by-point responses

  1. Referee: [Embedding-transcription consistency analysis] Embedding-transcription consistency analysis (abstract and corresponding results section): The strongest claim—that MuSViT encodes symbolic musical structure directly while other encoders do not—depends entirely on this analysis. No description is supplied of the alignment procedure between embeddings and transcribed content, the correlation or distance metric employed, whether transcription is performed by an independent symbol recognizer, or any controls that would isolate symbolic properties (pitch, rhythm, voice leading) from generic visual layout or glyph statistics. This detail is required to substantiate the claim.

    Authors: We agree the original manuscript lacked sufficient detail on this analysis. In revision we will add a dedicated subsection describing: (i) the alignment procedure, which projects MuSViT embeddings and independent OMR transcriptions into a shared space via linear probing on a held-out set of 50k pages; (ii) the metric, Pearson correlation between cosine distances in embedding space and normalized Levenshtein distances on the transcribed symbolic sequences; (iii) use of a separate, frozen OMR model (not trained on MuSViT data) for transcription; and (iv) controls that ablate layout statistics (via shuffled staff images) and glyph frequency (via bag-of-symbols baselines) to isolate pitch/rhythm/voice-leading correlations. These additions will allow direct evaluation of whether the observed correlations reflect symbolic structure. revision: yes

Circularity Check

0 steps flagged

No circularity: empirical pre-training and evaluation chain is self-contained

full rationale

The paper describes a standard empirical pipeline: pre-training a ViT encoder via masked autoencoders on the IMSLP corpus using a two-stage curriculum, followed by linear probing and fine-tuning evaluations on four downstream tasks plus an embedding-transcription consistency analysis. No equations, derivations, or parameter-fitting steps are presented that reduce any claimed performance or correlation result to quantities defined by the inputs themselves. The central claims rest on experimental outcomes rather than self-referential definitions, fitted-input predictions, or load-bearing self-citations. This is the most common honest finding for an empirical foundation-model paper.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Without the full manuscript, specific free parameters, axioms, or invented entities cannot be audited. The approach appears to rest on standard MAE pre-training assumptions and the representativeness of the IMSLP corpus for real-world sheet music.

pith-pipeline@v0.9.1-grok · 5810 in / 1207 out tokens · 32458 ms · 2026-07-01T06:10:31.074760+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

45 extracted references · 1 canonical work pages

  1. [1]

    InProceedings of the 42nd International Conference on Machine Learning, Vancouver, Canada, July 2025

    Audio Flamingo 2: An Audio-Language Model with Long-Audio Understanding and Expert Reasoning Abilities. InProceedings of the 42nd International Conference on Machine Learning, Vancouver, Canada, July 2025

    2025

  2. [2]

    Awais, M

    M. Awais, M. Naseer, S. Khan, R. M. Anwer, H. Cholakkal, M. Shah, M.-H. Yang, and F. S. Khan. Foundation Models Defining a New Era in Vision: A Survey and Outlook.IEEE Transactions on Pattern Analysis and Machine Intelligence, 47(4):2245–2264, 2025

    2025

  3. [3]

    J. Bai, S. Bai, Y . Chu, Z. Cui, K. Dang, X. Deng, Y . Fan, W. Ge, Y . Han, F. Huang, B. Hui, L. Ji, M. Li, J. Lin, R. Lin, D. Liu, G. Liu, C. Lu, K. Lu, J. Ma, R. Men, X. Ren, X. Ren, C. Tan, S. Tan, J. Tu, P. Wang, S. Wang, W. Wang, S. Wu, B. Xu, J. Xu, A. Yang, H. Yang, J. Yang, S. Yang, Y . Yao, B. Yu, H. Yuan, Z. Yuan, J. Zhang, X. Zhang, Y . Zhang, ...

    2023

  4. [4]

    S. Bai, Y . Cai, R. Chen, K. Chen, X. Chen, Z. Cheng, L. Deng, W. Ding, C. Gao, C. Ge, W. Ge, Z. Guo, Q. Huang, J. Huang, F. Huang, B. Hui, S. Jiang, Z. Li, M. Li, M. Li, K. Li, Z. Lin, J. Lin, X. Liu, J. Liu, C. Liu, Y . Liu, D. Liu, S. Liu, D. Lu, R. Luo, C. Lv, R. Men, L. Meng, X. Ren, X. Ren, S. Song, Y . Sun, J. Tang, J. Tu, J. Wan, P. Wang, P. Wang,...

    2025

  5. [5]

    Benesty, J

    J. Benesty, J. Chen, Y . Huang, and I. Cohen. Pearson correlation coefficient. InNoise reduction in speech processing, pages 1–4. Springer, 2009

    2009

  6. [6]

    Beyer, A

    L. Beyer, A. Steiner, A. S. Pinto, A. Kolesnikov, X. Wang, D. Salz, M. Neumann, I. Alabdul- mohsin, M. Tschannen, E. Bugliarello, T. Unterthiner, D. Keysers, S. Koppula, F. Liu, A. Grycner, A. Gritsenko, N. Houlsby, M. Kumar, K. Rong, J. Eisenschlos, R. Kabra, M. Bauer, M. Boˇsnjak, X. Chen, M. Minderer, P. V oigtlaender, I. Bica, I. Balazevic, J. Puigcer...

    2024

  7. [7]

    Calvo-Zaragoza, A

    J. Calvo-Zaragoza, A. H. Toselli, and E. Vidal. Handwritten Music Recognition for Mensural notation with convolutional recurrent neural networks.Pattern Recognition Letters, 128:115–121, 2019

    2019

  8. [8]

    Calvo-Zaragoza, J

    J. Calvo-Zaragoza, J. H. Jr, and A. Pacha. Understanding Optical Music Recognition.ACM Computing Surveys, 53(4):1–35, 2020

    2020

  9. [9]

    Caron, H

    M. Caron, H. Touvron, I. Misra, H. J ´egou, J. Mairal, P. Bojanowski, and A. Joulin. Emerging properties in self-supervised vision transformers. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 9650–9660, Online, June 2021. IEEE Computer Society. 12

    2021

  10. [10]

    Y . Chu, J. Xu, X. Zhou, Q. Yang, S. Zhang, Z. Yan, C. Zhou, and J. Zhou. Qwen-Audio: Advancing Universal Audio Understanding via Unified Large-Scale Audio-Language Models, 2023

    2023

  11. [11]

    Y . Chu, J. Xu, Q. Yang, H. Wei, X. Wei, Z. Guo, Y . Leng, Y . Lv, J. He, J. Lin, C. Zhou, and J. Zhou. Qwen2-Audio Technical Report, 2024

    2024

  12. [12]

    Dosovitskiy, L

    A. Dosovitskiy, L. Beyer, A. Kolesnikov, D. Weissenborn, X. Zhai, T. Unterthiner, M. Dehghani, M. Minderer, G. Heigold, S. Gelly, J. Uszkoreit, and N. Houlsby. An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale. InProceedings of the 9th International Conference on Learning Representations, Online, May 2021

    2021

  13. [13]

    Fujinaga

    I. Fujinaga. Staff detection and removal. InVisual Perception of Music Notation: On-Line and Off Line Recognition, pages 1–39. IGI Global Scientific Publishing, 2004

    2004

  14. [14]

    A. Goel, S. Ghosh, J. Kim, S. Kumar, Z. Kong, S.-g. Lee, C.-H. H. Yang, R. Duraiswami, D. Manocha, R. Valle, et al. Audio Flamingo 3: Advancing Audio Intelligence with Fully Open Large Audio Language Models. InProceedings of the 39th Annual Conference on Neural Informa- tion Processing Systems, Sydney, Australia, Dec. 2025

    2025

  15. [15]

    K. He, X. Chen, S. Xie, Y . Li, P. Doll´ar, and R. Girshick. Masked autoencoders are scalable vision learners. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 16000–16009, New Orleans, Louisiana, June 2022. IEEE Computer Society

    2022

  16. [16]

    Hondru, F

    V . Hondru, F. A. Croitoru, S. Minaee, R. T. Ionescu, and N. Sebe. Masked image modeling: A survey.International Journal of Computer Vision, 133(10):7154–7200, 2025

    2025

  17. [17]

    E. J. Hu, yelong shen, P. Wallis, Z. Allen-Zhu, Y . Li, S. Wang, L. Wang, and W. Chen. LoRA: Low-Rank Adaptation of Large Language Models. InInternational Conference on Learning Representations, Online, Apr. 2022

    2022

  18. [18]

    Huang, A

    Q. Huang, A. Jansen, J. Lee, R. Ganti, J. Y . Li, and D. P. W. Ellis. MuLan: A Joint Embedding of Music Audio and Natural Language. InProceedings of the 23rd International Society for Music Information Retrieval Conference, pages 559–566, Bengaluru, India, Dec. 2022. ISMIR

    2022

  19. [19]

    L. Jing, P. Vincent, Y . LeCun, and Y . Tian. Understanding dimensional collapse in contrastive self-supervised learning.arXiv preprint arXiv:2110.09348, 2021

    work page arXiv 2021

  20. [20]

    Kim, C.-B

    D. Kim, C.-B. Sohn, D.-Y . Kim, and D.-Y . Kim. A taxonomy and theoretical analysis of collapse phenomena in unsupervised representation learning.Mathematics, 13(18):2986, 2025

    2025

  21. [21]

    Z. Kong, A. Goel, R. Badlani, W. Ping, R. Valle, and B. Catanzaro. Audio Flamingo: A Novel Audio Language Model with Few-Shot Learning and Dialogue Abilities. InProceedings of the 41st International Conference on Machine Learning, Vienna, Austria, July 2024

    2024

  22. [22]

    Y . LI, R. Yuan, G. Zhang, Y . Ma, X. Chen, H. Yin, C. Xiao, C. Lin, A. Ragni, E. Benetos, N. Gyenge, R. Dannenberg, R. Liu, W. Chen, G. Xia, Y . Shi, W. Huang, Z. Wang, Y . Guo, and J. Fu. MERT: Acoustic Music Understanding Model with Large-Scale Self-supervised Training. InProceedings of the 12th International Conference on Learning Representations, Vie...

    2024

  23. [23]

    F. Luo, Y . Dai, J. Fuentes, W. Ding, and X. Zhang. M-DETR: multi-scale DETR for optical music recognition. volume 249, page 123664. Elsevier, 2024

    2024

  24. [24]

    T. Lv, Y . Huang, J. Chen, Y . Zhao, Y . Jia, L. Cui, S. Ma, Y . Chang, S. Huang, W. Wang, L. Dong, W. Luo, S. Wu, G. Wang, C. Zhang, and F. Wei. KOSMOS-2.5: A Multimodal Literate Model, 2025. 13

    2025

  25. [25]

    Madue˜no, A

    A. Madue˜no, A. R´ıos-Vila, and D. Rizo. Automatized incipit encoding at the Andalusian Music Documentation Center. InProceedings of the 8th International Conference on Digital Libraries for Musicology, Online, July 2021. Association for Computing Machinery

    2021

  26. [26]

    J. C. Martinez-Sevilla, D. Rizo, and J. Calvo-Zaragoza. Towards universal Optical Music Recogni- tion: A case study on notation types. InProceedings of the 25th International Society for Music Information Retrieval Conference, pages 914–921, San Francisco, USA, Nov. 2024. ISMIR

    2024

  27. [27]

    Oquab, T

    M. Oquab, T. Darcet, T. Moutakanni, H. V o, M. Szafraniec, V . Khalidov, P. Fernandez, D. Haziza, F. Massa, A. El-Nouby, M. Assran, N. Ballas, W. Galuba, R. Howes, P.-Y . Huang, S.-W. Li, I. Misra, M. Rabbat, V . Sharma, G. Synnaeve, H. Xu, H. Jegou, J. Mairal, P. Labatut, A. Joulin, and P. Bojanowski. DINOv2: Learning Robust Visual Features without Super...

    2024

  28. [28]

    Parada-Cabaleiro, A

    E. Parada-Cabaleiro, A. Batliner, and B. W. Schuller. A Diplomatic Edition of Il Lauro Secco: Ground Truth for OMR of White Mensural Notation. InProceedings of the 20th International Society for Music Information Retrieval Conference, pages 557–564, Delft, The Netherlands, Nov

  29. [29]

    Pugin, R

    L. Pugin, R. Zitellini, and P. Roland. Verovio: A library for Engraving MEI Music Notation into SVG. InProceedings of the 15th International Society for Music Information Retrieval Conference, pages 107–112, Taipei, Taiwan, Oct. 2014. ISMIR

    2014

  30. [30]

    Radford, J

    A. Radford, J. W. Kim, C. Hallacy, A. Ramesh, G. Goh, S. Agarwal, G. Sastry, A. Askell, P. Mishkin, J. Clark, et al. Learning Transferable Visual Models From Natural Language Supervision. In Proceedings of the 38th International Conference on Machine Learning, pages 8748–8763, Online, May 2021

    2021

  31. [31]

    Ramoneda, J

    P. Ramoneda, J. J. Valero-Mas, D. Jeong, and X. Serra. Predicting performance difficulty from piano sheet music images. InProceedings of the 24th International Society for Music Information Retrieval Conference, pages 708–715, Milan, Italy, Nov. 2023. ISMIR

    2023

  32. [32]

    Ramoneda, V

    P. Ramoneda, V . Eremenko, A. D’Hooge, E. Parada-Cabaleiro, and X. Serra. Towards explainable and interpretable musical difficulty estimation. InProceedings of the 25th International Society for Music Information Retrieval Conference, pages 520—-528, California, USA, Nov. 2024. ISMIR

    2024

  33. [33]

    S. Ren, K. He, R. Girshick, and J. Sun. Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks. InAdvances in Neural Information Processing Systems, volume 28, Barcelona, Spain, Dec. 2015. Curran Associates, Inc

    2015

  34. [34]

    R´ıos-Vila, M

    A. R´ıos-Vila, M. Espl`a-Gomis, D. Rizo, P. J. Ponce de Le´on, and J. M. I˜nesta. Applying Automatic Translation for Optical Music Recognition’s Encoding Step.Applied Sciences, 11(9):3890–3912, 2021

    2021

  35. [35]

    R´ıos-Vila, D

    A. R´ıos-Vila, D. Rizo, J. M. I˜nesta, and J. Calvo-Zaragoza. End-to-end optical music recognition for pianoform sheet music.International Journal on Document Analysis and Recognition, 26(3): 347–362, 2023

    2023

  36. [36]

    R´ıos-Vila, J

    A. R´ıos-Vila, J. Calvo-Zaragoza, D. Rizo, and T. Paquet. End-to-End Full-Page Optical Music Recognition for Pianoform Sheet Music.International Journal of Computer Vision, 134:49–66, 2026

    2026

  37. [37]

    Sim´eoni, H

    O. Sim´eoni, H. V . V o, M. Seitzer, F. Baldassarre, M. Oquab, C. Jose, V . Khalidov, M. Szafraniec, S. Yi, M. Ramamonjisoa, F. Massa, D. Haziza, L. Wehrstedt, J. Wang, T. Darcet, T. Moutakanni, L. Sentana, C. Roberts, A. Vedaldi, J. Tolan, J. Brandt, C. Couprie, J. Mairal, H. J´egou, P. Labatut, and P. Bojanowski. DINOv3, 2025. 14

    2025

  38. [38]

    Steiner, A

    A. Steiner, A. S. Pinto, M. Tschannen, D. Keysers, X. Wang, Y . Bitton, A. Gritsenko, M. Minderer, A. Sherbondy, S. Long, S. Qin, R. Ingle, E. Bugliarello, S. Kazemzadeh, T. Mesnard, I. Alab- dulmohsin, L. Beyer, and X. Zhai. PaliGemma 2: A Family of Versatile VLMs for Transfer, 2024

    2024

  39. [39]

    M. E. Thomae, J. E. Cumming, and I. Fujinaga. Digitization of Choirbooks in Guatemala. In Proceedings of the 9th International Conference on Digital Libraries for Musicology, pages 19–26, Prague, Czech Republic, July 2022. Association for Computing Machinery

    2022

  40. [40]

    Tuggener, Y

    L. Tuggener, Y . P. Satyawan, A. Pacha, J. Schmidhuber, and T. Stadelmann. DeepScoresV2, Sept. 2020

    2020

  41. [41]

    Tuggener, R

    L. Tuggener, R. Emberger, A. Ghosh, P. Sager, Y . P. Satyawan, J. A. Montoya-Zegarra, S. Gold- schagg, F. Seibold, U. Gut, P. Ackermann, J. Schmidhuber, and T. Stadelmann. Real World Music Object Recognition.Transactions of the International Society for Music Information Retrieval, 7 (1):1–14, 2024

    2024

  42. [42]

    P. Wang, S. Bai, S. Tan, S. Wang, Z. Fan, J. Bai, K. Chen, X. Liu, J. Wang, W. Ge, Y . Fan, K. Dang, M. Du, X. Ren, R. Men, D. Liu, C. Zhou, J. Zhou, and J. Lin. Qwen2-VL: Enhancing Vision-Language Model’s Perception of the World at Any Resolution, 2024

    2024

  43. [43]

    C. Wissler. The Spearman correlation formula.Science, 22(558):309–311, 1905

    1905

  44. [44]

    X. Zhai, B. Mustafa, A. Kolesnikov, and L. Beyer. Sigmoid loss for language image pre-training. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 11975–11986, Paris, France, Oct. 2023. IEEE Computer Society

    2023

  45. [45]

    Zhang, Y

    Q. Zhang, Y . Wang, and Y . Wang. How mask matters: Towards theoretical understandings of masked autoencoders.Advances in Neural Information Processing Systems, 35:27127–27139, 2022. 15 A Supplementary Material We provide additional details and results in this supplementary material, organized as follows: •Section A.1— Terminology glossary for non-music r...

    2022