Staff detection and removal using derivation and connected component analysis

摘要

Optical musical recognition is composed of a number of very significant phases of which one would be staff removal: one of the first phases within the system. This article focuses on this certain phase with the aim of recognizing staff lines and/or removing them for the purposes of producing some images only comprising of musical symbols. According to the procedure applied here, the in-image shall have to, first, be turned into the binary to, later, be compacted through the run-length coding technique of imaging. Finally, use has been made of the two properties: Staff Line Height and Staff Space Height along with the connected component technique so as to separate away primarily the staves from the musical scores thereof. The resulting out-image through the mentioned process has a series of a musical symbols (in addition to staff lines) like slur, dynamic wage, noise, and tie whose deletion has been carried out by the method of derivation. The derivation method is preformed in such a manner as to: first, the curvature on the part of the objects within images be obtained to: second, drop them upon the condition that their value should possibly exceed a certain limit.