Optical music recognition: state-of-the-art and open issues

Abstract

For centuries, music has been shared and remembered by two traditions: aural transmission and in the form of written documents normally called musical scores. Many of these scores exist in the form of unpublished manuscripts and hence they are in danger of being lost through the normal ravages of time. To preserve the music some form of typesetting or, ideally, a computer system that can automatically decode the symbolic images and create new scores is required. Programs analogous to optical character recognition systems called optical music recognition (OMR) systems have been under intensive development for many years. However, the results to date are far from ideal. Each of the proposed methods emphasizes different properties and therefore makes it difficult to effectively evaluate its competitive advantages. This article provides an overview of the literature concerning the automatic analysis of images of printed and handwritten musical scores. For self-containment and for the benefit of the reader, an introduction to OMR processing systems precedes the literature overview. The following study

Interpreting strokes on paper with a mobile assistant

“…For example, optical musical recognition for handwritten music has been available in commercial software. However, as Rebelo et al [20] report in a recent survey, there are still no robust methodologies to recognize handwritten musical scores.…”

Recognition of pen-based music notation with finite-state machines

“…That is, an image can be generated from pen strokes in order to make it pass through a conventional image-based system (offline recognition). Nevertheless, the performance of OMR systems is far from optimal, especially in the case of handwritten notation (Rebelo et al, 2012). Note that the main intention of a pen-based score 25 composition system is to provide musicians with an interface that is as friendly as possible.…”

“…To the best of our knowledge, and supported by relevant reviews such as those by Rebelo et al (2012) or Fornés & Sánchez (2014), no further research on the pen-based music recognition task has been carried out. As discussed 140 above, the solutions proposed to date are not satisfactory from a user point of view since the only way in which to recognize symbols is by following the rules proposed by each system.…”

Online recognition of handwritten music symbols

“…As information technology has been dramatically advanced during the last two decades, digitization deeply invades various fields of music such as preservation, duplication and distribution [1], and many composers and songwriters use digital devices and computer softwares nowadays. Although those programs provide the function of making and editing music scores using music instruments and computer input devices, pen and paper still occupy an important position among various composition tools.…”

“…The proposed method is trained and verified using the HOMUS dataset [12], which was recently released and is the largest online handwritten music symbol dataset. Figure 2 shows examples of twenty four symbols in a subset of the dataset used in this study 1 . Note that the intra-variation in each class (symbol) is very large.…”