Semi-Supervised Convolutive NMF for Automatic Piano Transcription
This work addresses the problem of automatic music transcription for piano, offering a semi-supervised approach that reduces data requirements, though it is incremental with generalization issues.
The paper tackles automatic piano transcription by proposing a semi-supervised convolutive NMF method that requires only one recording per note, showing it outperforms state-of-the-art low-rank factorization techniques but slightly underperforms supervised deep learning methods on the MAPS dataset.
Automatic Music Transcription, which consists in transforming an audio recording of a musical performance into symbolic format, remains a difficult Music Information Retrieval task. In this work, which focuses on piano transcription, we propose a semi-supervised approach using low-rank matrix factorization techniques, in particular Convolutive Nonnegative Matrix Factorization. In the semi-supervised setting, only a single recording of each individual notes is required. We show on the MAPS dataset that the proposed semi-supervised CNMF method performs better than state-of-the-art low-rank factorization techniques and a little worse than supervised deep learning state-of-the-art methods, while however suffering from generalization issues.