Instrument Classification of Solo Sheet Music Images

Abstract

This paper studies instrument classification of solo sheet music. Whereas previous work has focused on instrument recognition in audio data, we instead approach the instrument classification problem using raw sheet music images. Our approach first converts the sheet music image into a sequence of musical words based on the bootleg score representation, and then treats the problem as a text classification task. We show that it is possible to significantly improve classifier performance by training a language model on unlabeled data, initializing a classifier with the pretrained language model weights, and then finetuning the classifier on labeled data. In this work, we train AWD-LSTM, GPT-2, and RoBERTa models on solo sheet music images from IMSLP for eight different instruments. We find that GPT-2 and RoBERTa slightly outperform AWD-LSTM, and that pretraining increases classification accuracy for RoBERTa from 34.5% to 42.9%. Furthermore, we propose two data augmentation methods that increase classification accuracy for RoBERTa by an additional 15%.

Towards the Generation of Musical Explanations with GPT-3

“…One which has taken the interest of researchers is the use of transformer-based language models for different musical-related tasks. For instance, previous work compared the performance of different transformer-based approaches for composer style classification [23] and instrument classification [16] by converting raw images of piano sheet music to text and treating these problems as text classification tasks. In [2], the XLNet language model is used to recognise emotion from lyrics, while in [11] GPT-2 is used to generate musical sequences in ABC notation.…”