Detecting Arabic handwritten word baseline using Voronoi Diagram

Abstract

The purpose of this paper is to present a methodology based on Voronoi Diagram for detecting the Arabic handwritten word baseline. In this paper, the definition and background of Voronoi Diagram has been explained. A short survey about the application of Voronoi Diagram on the language recognition fields also presented. Finally, a method is proposed based on the Voronoi Diagram for detecting the Arabic baseline handwritten word. The proposed method expects to be the robust Arabic handwriting words baseline detection algorithms, because the images slope and the diacritics will not affect in the method performance. In addition it will be the first method for detecting the Arabic baseline in zigzag way. The proposed method will implement in the next phase, in order to build a complete Arabic optical character recognition system.

Novel Algorithm for Baseline Detection of Offline Arabic Handwritten Text Recognition

“…However, the accuracy of this method is still affected by problems like skew, unbalanced size of the characters, and short words. Al-Shatnawi [37] used voronoi diagrams to estimate the straight and curved Arabic baseline. Depending on the clustering of interest points, Fawzi et al, [20] presented a method that operated by training and testing Arabic text database.…”

Untitled

“…Thereafter, the detection methods grew more and more sophisticated and researchers started using various techniques like the word skeleton (e.g., Pechwitz and Maergner [29]), principal component analysis (e.g., Burrow [13]), contour representation (e.g., Farooq et al [15]), and VDs (e.g., Al-Shatnawi [10]). Additional details on the various baseline detection methods have been provided by Al-Shatnawi and Omar [9], Al-Shatnawi and Omar [2], and Al-Shatnawi [10].…”

A hybrid Feature Selection Approach for Arabic Handwritten Text Based on Genetic Algorithm and Support Vector Machine

“…The Arabic language is a universal language and the formal language of 25 countries and more than 300 million persons worldwide. Furthermore, Arabic characters are used in several languages like Jawi, and Urdu languages [5], [1], [8]. Review of the literature reveals that there are two main systems for offline Arabic text recognition; segmentation-based systems and segmentation-free systems (holistic recognition approaches [12], [16].…”