Open AccessApplication of Artificial Neural Networks Technology for Handwritten Arabic Letters Recognition
Author(s) -
Aisha Douma,
Abdussalam Ali Ahmed,
Abdulgader Alsharif,
Mohamed Belrzaeg
Publication year - 2022
Publication title -
international journal of emerging trends in engineering research
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.218
H-Index - 14
ISSN - 2347-3983
DOI - 10.30534/ijeter/2022/161022022
Subject(s) - computer science , artificial neural network , handwriting , artificial intelligence , intelligent character recognition , arabic , pattern recognition (psychology) , handwriting recognition , speech recognition , similarity (geometry) , pixel , image (mathematics) , feature extraction , character recognition , linguistics , philosophy
The mechanism that convert and detect the handwriting letters by using machine-encoded forms is called handwriting recognition. The interaction between machines and humans is very important so that the handwriting recognition must be found. The complexity of Arabic letters and the similarity of at least three letters is main challenge to recognize them .Thus, our main challenge is to propose a methodology, implementation and evaluation of Arabic letters recognition system by using Artificial Neural Network approach in order to achieve high accuracy with some techniques will be produced in this paper. In this paper, we apply recognition artificial neural network (ANN) for Arabic letters. We use the intensity values of pixels for input of the neural network. These results show that ANN with high number of training images have the highest performace.