MUSCAT: a Multimodal mUSic Collection for Automatic Transcription of real recordings and image scores
About • How To Use • Citations • Acknowledgments • License
This repository is a modification of the Alfaro-Contreras et al. work: "A Transformer Approach For Polyphonic Audio-To-Score Transcription". It pretends to serve as an initial transcription baseline to the real music audio dataset MUSCUTS.
This dataset is a subset of the MUSCAT dataset. Each sample is an approximately 30-second cut from a longer song. The dataset can be downloaded and more information can be found at the following link.
The basic requirements are specified in the Dockerfile and requirements.txt as docker has been used. However, it is not mandatory, as any other virtual environment can be created.
The experiments were carried out with an initial training phase, like the script train_muscat; and a final testing phase providing the previous best weights as the script test_muscat.
@inproceedings{
galan-cuenca2024muscat,
title={{MUSCAT}: a Multimodal m{US}ic Collection for Automatic Transcription of real recordings and image scores},
author={Alejandro Galan-Cuenca and Jose J. Valero-Mas and Juan C. Martinez-Sevilla and Antonio Hidalgo-Centeno and Antonio Pertusa and Jorge Calvo-Zaragoza},
booktitle={ACM Multimedia 2024},
year={2024},
url={https://openreview.net/forum?id=B3CsOcxXOa}
}
This work is part of the I+D+i PID2020-118447RA-I00 (MultiScore) project, funded by MCIN/AEI/10.13039/501100011033.
This work is under a MIT license.