Sale!
Placeholder

A Survey of Optical Music Recognition Software Steven Driver

10,000 3,000

Topic Description

Summary
Optical Music Recognition (OMR) is a method for the automatic inputting of Sheet Music into a
machine-readable format, allowing the information to be used by the computer for tasks such as
editing or MIDI sequencing. The objective of any OMR system is to automate this process and hence
to save the user time by removing the need to do this manually. A number of Optical Music
Recognition programs have been created with this in mind, claiming varying levels of success.
This is a survey of commercially available Optical Music Recognition Software and is intended to
assess the performance of these packages by benchmark testing using a standard data set and to
address any problems that arise with the software. No real survey of this type has been made before
and it is not a simple thing to perform. There are many problems associated with recognition such as
interconnection problems and deciding what criteria to assess in order to give an accurate and useful
conclusion.
This is a report of such a survey and it discusses these factors and the formulation of a test plan based
on representative benchmarking data including discussion of criteria for selection. This is followed by
an analysis of the results to determine accuracy measures for each program tested and to assess the
most common problems with these packages

Contents
Summary……………………………………………………………………………………………………………………. i
Acknowledgements ………………………………………………………………………………………… ii
Contents ……………………………………………………………………………………………………………………. iii
1. Introduction……………………………………………………………………………………………………… 2
1.1 Project Aims ………………………………………………………………………………………………………………… 2
1.2 Objectives ……………………………………………………………………………………………………………………. 2
1.3 Minimum Requirements ………………………………………………………………………………………………… 3
2. Background……………………………………………………………………………………………………… 4
2.1 Previous Surveys ………………………………………………………………………………………………………….. 4
2.2 Input……………………………………………………………………………………………………………………………. 5
2.3 Pre-processing ……………………………………………………………………………………………………………… 6
2.4 Segmentation ……………………………………………………………………………………………………………….. 7
2.4.1 Graphical Primitives ……………………………………………………………………………………………….. 8
2.5 Recognition………………………………………………………………………………………………………………….. 9
2.5.1 k-Nearest-Neighbour ………………………………………………………………………………………………. 9
2.5.2 Genetic Algorithms…………………………………………………………………………………………………. 9
2.6 Musical Semantics………………………………………………………………………………………………………. 10
2.7 Output ……………………………………………………………………………………………………………………….. 10
3. OMR Survey…………………………………………………………………………………………………. 11
3.1 OMR Packages …………………………………………………………………………………………………………… 11
3.2 Test Data……………………………………………………………………………………………………………………. 11
3.2.1 Criteria for Selection …………………………………………………………………………………………….. 12
3.2.2 Ground Truth Data Set…………………………………………………………………………………………… 14
3.3 Assessment Criteria …………………………………………………………………………………………………….. 14
3.4 Testing ………………………………………………………………………………………………………………………. 15
3.5 Symbol Density ………………………………………………………………………………………………………….. 16
4. Analysis …………………………………………………………………………………………………………….. 17
4.1 Results of Assessment …………………………………………………………………………………………………. 17
4.1.1 SmartScore…………………………………………………………………………………………………………… 17
4.1.2 Music Publisher Scan ……………………………………………………………………………………………. 19
4.1.3 Photoscore Professional…………………………………………………………………………………………. 21
4.2 Conclusions ……………………………………………………………………………………………………………….. 23
iv
5. Evaluation……………………………………………………………………………………………………….. 26
5.1 Evaluation Criteria………………………………………………………………………………………………………. 26
5.2 Results of Evaluation…………………………………………………………………………………………………… 27
5.2.1 Aim and Minimum Requirements …………………………………………………………………………… 27
5.2.2 Timekeeping ………………………………………………………………………………………………………… 27
5.2.3 Methodology………………………………………………………………………………………………………… 29
5.2.4 Report Structure……………………………………………………………………………………………………. 29
5.3 Conclusions ……………………………………………………………………………………………………………….. 29
6. Extensions and Future Directions……………………………………………… 31
6.1 Possible Extensions …………………………………………………………………………………………………….. 31
6.2 Future Directions ………………………………………………………………………………………………………… 32
Bibliography………………………………………………………………………………………………………… 33
Appendix A – Project Reflection……………………………………………………… 36
Appendix B – Ground Truth Data Set

GET COMPLETE PROJECT