DNN-based Arabic Printed Characters Classification

The recognition of Arabic characters in scanned images, handwritten and printed texts is a difficult operation that relies on Optical Character Recognition (OCR). The Arabic OCR scores are dismal, but this is still an active study topic. High recognition accuracy and minimal processing costs are achieved by utilizing the best feature extraction techniques and a suitable classification algorithm. In this study, we created a model for recognizing printed Arabic characters using Deep Learning (DL). After evaluating the recognition system, the Arabic printed character recognition (APCR) results showed an accuracy of 99.88%.