Music scores have a different typical use case from printed books in that many scores representing music for which copyright has expired, but such music is still widely consumed.
Because such music is widely consumed, new scores for old music are often published. While the copyright of the music has expired, the copyright of the engraving (this is what it is called when scores are "typeset") is recent, allowing the publishers to charge considerable amounts for these scores.
The purpose of this project is to create an interactive system that allows an operator to processed a scanned version of a music score for which the copyright for both the music and the engraving has expired.
Automatically recognizing music from a scanned score is known to be a hard problem, and no universal solutions exist. Instead, the proposed project would result in a system that will be used for assisting a human operator in identifying the various symbols on the score. The result wold be a set of pairs of symbols and positions, where the symbols are recovered from the scanned score. There would also be other information that is not in the form of symbols, such as staves, ties, and slurs.
A possible interface that would work sufficiently well for small scores would be to make the user mark each symbol with an approximate location using the mouse. Once could for instance start by drawing a bounding box around all staves, then around all clefs. Note heads would be marked by a simple click with the mouse. A symbol that is recognized by the system changes color so that unrecognized symbols stand out.