Character segmentation based on the Arabic script

Abstract

Character recognition for Arabic script-based languages using segmentation has long been a major area of study. Researchers from the academic and industrial worlds have focused on the complex nature of Arabic script recognition, but their efforts have not yet shown promising results. Due to the writing style's intricacy compared to Naskh writing style, segmenting Urdu script produced in Nasta'liq is a particularly challenging process. One of the factors in high accuracy is effective segmentation. The character segmentation stage of the OCR process has proved crucial. The value of segmentation is clearly demonstrated by the better identification rates for solitary characters as compared to findings for words or related characters. The current paper examines recent character segmentation efforts.