This code is a set of tools for doing preprocessing of sheetmusic with gamera and the gamera toolkit musicstaves.
The purpose of this preprocessing is to eliminate anything that might confuse current commercial Optical Music Recognition systems like Photoscore or Sharpeye.
This code is developed as a part of a project named Optical character recognition for structural information from high-quality scanned music which is part of the study credit for my two-years masters(cand.scient) at the University Of Copenhagen, Department of Computer Science
Other parts can be seen in
- MusicCrawler github - Code that uses BeautifulSoup to crawl sheet music libraries for available sheet music.
- Diku ONR Github. LaTeX source of the paper "driving" this code. (Work In Progress)
-
class_dynamic.py
$ python class_dynamic.py mergedyn2.xml 0.99 lvbp0.tif
First argument is the training data, second is e_fp and last is a list of filenames
-
mytest.py
$ python mytest.py lvbp0.tif 8
-
proj.py
$ python proj.py lvbp0.tif
As Gamera is GPL this is also GPL v2. See LICENSE.txt for more info