Recognizing Composer’s Musical Signature from Bars of Music Using Computer Vision and Deep Learning

Abstract

For centuries, music has been divided over two traditions in the form of written documents and aural transmission normally called musical scores. Many of these scores are published in manuscript form and So they are at risk of being lost over time. The system takes a music score image as input, segments music symbols after preprocessing the image, then recognizes their pitch and duration. Finally, MIDI files are generated. Similar to optical music recognition (OMR) systems, programs similar to optical character recognition systems have been in intensive development for many years. This thesis provides an overview of automatic analysis of handwritten music scores. An overview of the literature provides an overview of OMR processing systems for the benefit of the reader and self-interest. The OMR system can provide various benefits to the scientific community for an effective and powerful printed and handwritten music score. We have presented some of the strategies of OMR utilizing computer vision and deep learning based approaches for music scores and printed processing.