Notice of Violation of IEEE Publication Principles: Music Staff Elimination using Supervised Pixel Classification
Annadanesh Javaji, Santosh Chinchali, Shivakumar Honawad · 2020 IEEE Bangalore Humanitarian Technology Conference (B-HTC) · 2020
This paper presents an approach for music staff elimination from an image containing music score without disturbing any symbol information. This is one of the most important tracks which improve the realization in optical music recognition system. Depending on intrinsic of music scores staff elimination was performed in literature based on image processing techniques. Here, a problem is modeled as a supervised pixel learning classification task for the elimination of staff lines. In this scenario, front pixel is tagged as symbol or staff. To train the classification algorithms, a pairs of scores with and without stafflines are used. Our proposed methodology is tested with other well-know classification techniques. Moreover, in our experiment certain parameters are set to default setting which is provided by libraries, an attempt of tuning the classification algorithm is not permitted. Our ultimate aim represents that even though by applying the straightforward method, still our method gives competitive results by using significant algorithms. Some of the advantages of using this method over the earlier methods are its high versatility.