ORPHEUS
Optical Recognition Parser of Harmonization with Edition Utilities for Scores
Business goal: Create an accessible and accurate OMR application that automates the conversion of printed sheet music into an editable digital format. Improve the efficiency of music transcription, editing, and playback while lowering entry barriers for musicians and educators. Outcomes:
- Development of the desktop application integrating an OMR module (Python, PyTorch) with an editor (C#, WPF).
- Reliable MusicXML import, editing, and playback available in one environment.
- A trained OMR model achieving over 70% symbol recognition accuracy.
- Fully integrated, tested, and stable MVP ready for user evaluation and presentation. Verification:
- Model accuracy measured on test datasets.
- Editor usability verified through user testing.
- System stability and performance validated through integration and functional tests.
Transcribing printed sheet music into digital form is a slow, manual, and error-prone process. Existing Optical Music Recognition (OMR) tools are often expensive, unreliable on imperfect scans, and lack integration with music editing software. There is a need for an accessible, accurate, and unified solution that allows users to scan, digitize, and edit musical scores seamlessly within a single environment.
Final deliverable: Desktop application integrating Optical Music Recognition with an interactive music editor, enabling users to scan, edit, and play sheet music. Supporting tools: Training scripts for model development and testing. Characteristics: Developed in Python (OpenCV, PyTorch) and C# (WPF, Manufaktura Controls). Runs locally on Windows, performing image processing, recognition, MusicXML generation, editing, and playback within one environment.
End users: Musicians of all kinds — composers, arrangers, performers, and educators — will save time on transcription, edit scores more easily, and use playback to improve learning and rehearsal efficiency. External organisations: Music schools and cultural institutions can integrate ORPHEUS into teaching or archiving workflows, improving accessibility and modernisation of their resources. Communities: The project supports the broader music community by making digital score editing and analysis more accessible, encouraging learning and collaboration.
