# Composer Style Classification of Piano Sheet Music Images Using Language Model Pretraining

arXiv: https://arxiv.org/abs/2007.14587 · HTML: https://mlsanity.com/paper/2007.14587/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 29 Jul 2020.

## Built on by (most cited first, 7 of 7)

- [DadaGP: A Dataset of Tokenized GuitarPro Songs for Sequence Models](https://mlsanity.com/paper/2107.14653/) (2021)
- [BERT-like Pre-training for Symbolic Piano Music Classification Tasks](https://mlsanity.com/paper/2107.05223/) (2021)
- [PianoBART: Symbolic Piano Music Generation and Understanding with Large-Scale Pre-Training](https://mlsanity.com/paper/2407.03361/) (2024)
- [Towards the Generation of Musical Explanations with GPT-3](https://mlsanity.com/paper/2206.08264/) (2022)
- [Proceedings of the 4th International Workshop on Reading Music Systems](https://mlsanity.com/paper/2211.13285/) (2022)
- [PBSCR: The Piano Bootleg Score Composer Recognition Dataset](https://mlsanity.com/paper/2401.16803/) (2024)
- [Instrument Classification of Solo Sheet Music Images](https://mlsanity.com/paper/2609.18980/) (2026)

## Abstract (arXiv)

This paper studies composer style classification of piano sheet music images. Previous approaches to the composer classification task have been limited by a scarcity of data. We address this issue in two ways: (1) we recast the problem to be based on raw sheet music images rather than a symbolic music format, and (2) we propose an approach that can be trained on unlabeled data. Our approach first converts the sheet music image into a sequence of musical "words" based on the bootleg feature representation, and then feeds the sequence into a text classifier. We show that it is possible to significantly improve classifier performance by first training a language model on a set of unlabeled data, initializing the classifier with the pretrained language model weights, and then finetuning the classifier on a small amount of labeled data. We train AWD-LSTM, GPT-2, and RoBERTa language models on all piano sheet music images in IMSLP. We find that transformer-based architectures outperform CNN and LSTM models, and pretraining boosts classification accuracy for the GPT-2 model from 46% to 70% on a 9-way classification task. The trained model can also be used as a feature extractor that projects piano sheet music into a feature space that characterizes compositional style.