Off-line Optical Character Recognition System for Arabic Handwritten text

Mohamed A. Fadeel (1)
(1) , Libya

Abstract

This paper introduces an Optical Character Recognition (OCR) system, where different stages; pre-processing, thinning, segmentation, features extraction, classification and recognition were designed, tested and implemented in the overall system. Learning Vector Quantization algorithm was first used as classifier for Arabic handwritten characters recognition. Two classification strategies were performed; one classifier for all Arabic Alphabets and three classifier (one for ascenders, one for descenders and one for embedded). The later strategy was adopted in classification stage of the proposed system. Numbers of experiments were run on different samples of Arabic handwritten scripts for different writers, a very satisfactory recognition accuracy was obtained.

Full text article

Generated from XML file

Authors

Mohamed A. Fadeel
Mohamed A. Fadeel. (2019). Off-line Optical Character Recognition System for Arabic Handwritten text. Journal of Pure & Applied Sciences, 18(4). https://doi.org/10.51984/jopas.v18i4.384

Article Details

No Related Submission Found