Robust and Adaptive OMR System Including Fuzzy Modeling, Fusion of Musical Rules, and Possible Error Detection

Abstract

This paper describes a system for optical music recognition (OMR) in case of monophonic typeset scores. After clarifying the difficulties specific to this domain, we propose appropriate solutions at both image analysis level and high-level interpretation. Thus, a recognition and segmentation method is designed, that allows dealing with common printing defects and numerous symbol interconnections. Then, musical rules are modeled and integrated, in order to make a consistent decision. This high-level interpretation step relies on the fuzzy sets and possibility framework, since it allows dealing with symbol variability, flexibility, and imprecision of music rules, and merging all these heterogeneous pieces of information. Other innovative features are the indication of potential errors and the possibility of applying learning procedures, in order to gain in robustness. Experiments conducted on a large data base show that the proposed method constitutes an interesting contribution to OMR.

A new optical music recognition system based on combined neural network

“…S. Cardoso et al, 2009;Ana Rebelo and Jaime S. Cardoso, 2013;C. Dalitz, 2008), music symbol segmentation(F. Rossant and I. Bloch, 2007;Forns et al, 2005) and music recognition system approaches(G. S. Choudhury et al, 2000). Recently, (Ana Rebelo et al, 2013) proposed a parametric model to incorporate syntactic and semantic music rules after a music symbols segmentation's method.…”

A Robust Staff Line Height and Staff Line Space Estimation for the Preprocessing of Music Score Recognition

“…Music sheet image usually suffer degradation such as curve or skew therefore horizontal projection approach does not work efficiently in real cases. Other methods are proposed to overcome degradation such as using Hough Transform to detect staff lines [9], Line Adjacency Graph (LAG) [10][11][12]; line tracing [13][14][15]. Fujinaga et al in [6] use many different techniques such as run-length coding, connected-component analysis and projection for finding staff lines.…”

Music staff removal with supervised pixel classification

“…This strategy was improved and extended to be used on grey-scale scores [27]; Dutta et al [10] developed a method that considers the staff line segment as a horizontal connection of vertical black runs with uniform height, which are validated using neighbouring properties; in the work of Piatkowska et al [21], a swarm intelligence algorithm was applied to detect the staff line patterns; Su et al [31] start estimating properties of the staves like height and space; then, they tried to predict the direction of the lines and fitted an approximate staff, which was posteriorly adjusted; Geraud [13] developed a method that entails a series of morphological operators directly applied to the image of the score to remove staff lines; and Montagner et al [19] proposed to learn image operators, following the work of Hirata [17], whose combination was able to remove staff lines. Others works have addressed the whole OMR problem by developing their own, case-directed staff removal process [29,32].…”