A fast, reliable sheet music informatics system
python show_layout.py samples/sonata.png
This detects layout-related information from the score (staves, measures, and staff systems) and overlays the result with the image using matplotlib.
- Robust layout (staves, systems, and measures) detection
- Random forest glyph classification
- Identifying multiple scans and editions of the same piece, and automatically selecting higher-quality scans
- Creating a high-level OMR accuracy benchmark
- Python 2.7+
- NumPy
- Reikna
- PyOpenCL or PyCUDA
- PyFFT
- Pillow, PyPDF, pylibtiff (PDF loading)
- Pylab (for displaying data only)
- SciPy (for Kanungo noise parameter estimation)
- Music21 (for classified note export)