|
Acadêmico(a): Jonathan Mauricenz |
Título: Interpres: um Protótipo para Reconhecimento de Partituras |
|
Abstract: |
This paper describes the development of a prototype in Java can recognize musical symbols from a picture score. For the classification of musical symbols became necessary to remove the staff lines for horizontal projection of black pixels, isolate the musical symbols using mathematical morphology, calculate the Fourier descriptors to characterize the shape and decision tree as a method of classification. The results show that the prototype has a rate of 77% accuracy.
|
|
|