Automating Building of Optical Character Recognition (OCR) Based on Web Services

Abstract

Optical Character Recognition (OCR) is a computer system designed to transform images of typewritten text (typically captured by a scanner) into machine-processed text. Arabic OCR has been developed and enhanced over decades leading to the presence of the enormous number of approaches with robust results that approaching accuracy, in some cases, of approximately 99%. However, existing OCRs exhibit shortages, or at least only sub-sets of them, when they were implemented with new applications, such as low-resolution inputs. In this thesis, we will improve results Optical Character Recognition (OCR) by choose our original image based on (APTI) Arabic Printed Text Image Database. Next, we will improve inputs by processing the images to make them clear and ready to be inserted into the OCR. Then, we will test all Arabic OCR websites online to find the best one which currently exists. Finally, we will create web service that can optimize image effects automatic and insert it to best the Arabic OCR to make the