Multiform frame shape extraction using neighbouring pixel be-haviors for Mushaf Al-Quran

Abstract

Page decoration removal in the Al-Quran is an important step in document processing. One of the challenging parts of Al-Quran text recognition area is text extraction from its decorated page. Al-Quran decorations contain different patterns and textures depending on the Mushaf. In this paper, we propose a novel approach based on neighbouring properties by analysing the binary image page of current neighbouring pixels of point. This paper will focus on how to extract Al-Quran text from its decorative illumination frame. The accuracy of the result of the proposed algorithm depends on the noise of the images. The proposed method may contribute to the development of applications related to Optical Character Recognition (OCR).