On the Potential of Fully Convolutional Neural Networks for Musical Symbol Detection
Matthias Dorfer, Jan Hajič, Gerhard Widmer · 2017
Musical symbol detection on the page is an outstanding Optical Music Recognition (OMR) subproblem. We propose using a fully convolutional segmentation network to produce high-quality pixel-wise symbol probability masks. Experiments on notehead detection show a very promising detection f-score of 0.98 with elementary detection methods.