Review on Arabic Handwritten Recognition Using Deep Learning and Machine Learning

Abstract

Recognizing Arabic handwriting using deep learning and machine learning is a contemporary technique in digital image processing that has potential applications ranging from recognition and indexing of Arabic texts to conversion into electronic texts, and enhancing the user experience on smart devices and computers. the unique and Arabic's use of cursive writing presents a challenge for recognition systems when it comes to identifying handwritten characters, digits, and words. This paper aims to discuss the challenges of recognizing Arabic handwriting and present various approaches to address them. The research involved a comprehensive survey of 40 research papers related to Arabic handwritten recognition, covering a wide range of topics related to the subject. The team compared the accuracy achieved by different methods used in these papers, proposed future work, and the datasets used. Their findings reveal that the highest accuracy achieved was 99.88% using the k-NN algorithm in combination with filters such as HOG, LBP, and Gabor. This approach offers a thorough overview of the cutting-edge in machine learning and deep learning for Arabic handwritten recognition. This study concludes that applying machine learning and deep learning in recognizing Arabic handwriting can achieve high levels of accuracy and can be used to improve user experience and increase the efficiency of document processing. The main recommendation is to further develop algorithms and techniques that use deep learning and machine learning in recognizing Arabic handwriting. This study's findings can be utilized to develop and enhance prove Arabic handwriting recognition systems, as well as to enhance interest and identify future directions for research and development in this field.