Automatic identification of music notations

The music scores of the 18th century and earlier were produced only in a manual way. The important problem of the current registration of old historical music scores lies in the identification of the corresponding writer to estimate who did the notation writing or drawing. We are developing efficient identification of a writer's hand. This identification is based on the automated image analysis approach. The theoretical manual identification is being developed at the Institute for Musicology at the University of Rostock. First tests of an analogue approach were applied on chosen music notations. These tests indicate that the recognition problem can be sufficiently solved only by computer-aided assistance. Our research activities are concentrated on the development of an automated identification of writers by analysing the notation graphic features. In our article we describe an approach for the automated image analysis and understanding as applied to digitised music scores. This approach is aimed at solving the automated identification of writers as a result of digital music notation matching.