Abstract
This article introduces a method for large-scale retrieval of piano sheet music images. We study this problem in two different scenarios: camera-based sheet music identification and MIDI-sheet image retrieval. Our proposed method combines bootleg score features with a novel hashing scheme called dynamic N-gram fingerprinting. This hashing scheme ensures that every fingerprint is discriminative enough to warrant a table lookup, which improves both retrieval accuracy and runtime. On experiments using all piano sheet music images in the IMSLP database, the proposed method achieves >0.8 mean reciprocal rank with sub-second runtimes. As a practical application, we use our system to find matches between the Lakh MIDI dataset and IMSLP, which augments the IMSLP sheet music data with symbolic music information for a subset of pieces. We release our code and Lakh-IMSLP matches to facilitate future study.
Highlights
This article investigates large-scale retrieval of piano sheet music images through two applications
To understand how our work fits into the broader landscape, we provide an overview of previous work on retrieval of sheet music images
All systems use a pretrained CNN as a backbone, but employ different approaches to reduce the activations to a fixedlength feature representation. Because these four retrieval systems were trained on natural images and not sheet music, we expect them to perform poorly
Summary
This article investigates large-scale retrieval of piano sheet music images through two applications. The first application is camera-based sheet music identification, where a user can take a cell phone picture of a physical page of piano sheet music and identify the matching sheet music in a large database. We will use our approach to identify matches between the Lakh MIDI Dataset (LMD) (Raffel, 2016) and IMSLP. This is an extremely large-scale, cross-modal retrieval task. In both applications, our work lays a foundation for searching large databases of sheet music in novel ways
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callMore From: Transactions of the International Society for Music Information Retrieval
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
