Optical music recognition : feature identification (1995)
Although it has been less than a year since the last progress report, work has reached a natural 'breakpoint' offering an opportunity to describe what has been accomplished as well as gathering thoughts on the future. For a broader picture of how the implemented work fits into the general design, the reader is directed towards [Bai94a], which describes a complete Optical Music Recognition system. Since it is possible for a piece of music to include arbitrary graphics1 , it is impossible to design an OMR system that can process all music. The key idea, therefore, expressed in [Bai94a] is to provide a versatile foundation that can be built upon by individual users to generate particular instances of the system capable of recognising a particular class of music notation. Such a philosophy is reminiscent of Computer Aided Design (CAD). Similar to this area, the user should be encouraged to utilise sound software engineering principles. The proposed system could be thought of as a Computer Aided Music Recognition (CAMR). The main body of this report describes the implemented work. The topics: staff separation; primitive identification; primitive data tabulation; drawing package development; prototype2 musical feature classifier extensions; and miscellaneous items are discussed in turn. The report concludes by reflecting on the completed work as well as contemplating how the remaining problems may be solved. Having the right development environment to study OMR is considered an important facet of work. Time throughout the project has been devoted to this. In addition to the topics listed in (Bai94b], time has recently been invested learning Perl ( a Practical Extraction and Report Language), Gofer ( a functional language), and Lime ( a music file format for graphical reconstruction). Developing Internet searching skills using 'whois', 'archie', 'gofer', and 'mosaic' has also proved invaluable. Finally, in lieu of the forth coming design of a musical knowledge expression language, time has been spent investigating the suitability of Lisp, Prolog and various public domain knowledge based systems. The excerpt of music shown in Figure 1 is a recurring example used, throughout this report, to illustrate various points.
ANZSRC Fields of Research46 - Information and computing sciences::4603 - Computer vision and multimedia computation::460306 - Image processing
08 - Information and Computing Sciences::0801 - Artificial Intelligence and Image Processing::080104 - Computer Vision
RightsAll Rights Reserved
Showing items related by title, author, creator and subject.
Bainbridge, David (University of Canterbury. Computer Science and Software Engineering, 1994)The purpose of writing this report is to record and comment on the work done over the last year. The report will also summarise my main insights into the problem and outline future work. There were three main areas of ...
Bainbridge, David (University of Canterbury. Computer Science and Software Engineering, 1994)Reading music is something a child can learn, and once understood, it becomes such a natural process that it is no longer a conscious effort. If we were to dissect this `natural process,' we might hypothesise that reading ...
Clark, Adrian (2018)Abstract: An aspect of the invention provides a texture mapping system configured to map texture to a three-dimensional model. The system includes a video renderer configured to render captured image data on a display to ...