Arabic (Indian) digit handwritten recognition using recurrent transfer deep architecture
- PDF / 1,106,361 Bytes
- 11 Pages / 595.276 x 790.866 pts Page_size
- 118 Downloads / 340 Views
METHODOLOGIES AND APPLICATION
Arabic (Indian) digit handwritten recognition using recurrent transfer deep architecture Rami S. Alkhawaldeh1
© Springer-Verlag GmbH Germany, part of Springer Nature 2020
Abstract The rapid volume of digit texts and images motivates researchers to build solid and efficient prediction models to recognize such media. The Arabic language is considered one of the difficult languages regarding the way of writing characters and digits. Recent research focuses on such language for building predictive approaches to recognize written materials. The Arabic (Indian) digit recognition task has been a challenging task and has gained more attention from researchers who build optimal predictive models from historical images that are used in many applications. However, transfer learning approaches exploit deep pre-trained models that could be re-used for similar tasks. So, in this paper, we propose an adapted deep hybrid transfer model developed using two well-known pre-trained convolutional neural networks (CNN) models. These are further adapted by adding recurrent neural networks especially long short-term memory (LSTM) architectures to detect Arabic (Indian) Handwritten Digits (AHD). The CNN model learns the relevant features of Arabic (Indian) digits, while the sequence learning process in the LSTM layers extracts long-term dependence features. The experimental results, using popular datasets, show significant performance obtained by the adapted transfer models with accuracy reached up to 98.92% as well as with precision and recall values at most cases reached to 100% with statistical t test using p-value ( p ≤ 0.05) compared to baseline methods. Keywords Arabic (Indian) handwritten recognition · Deep supervised learning · LSTM · Transfer learning
1 Introduction Handwritten digit recognition (HDR) is being one of the attractive areas in the fields of image processing and pattern recognition (Abdleazeem and El-Sherif 2008; AlKhateeb and Alseid 2014; Loey et al. 2017; Mudhsh and Almodfer 2017; Dargan et al. 2019; Elleuch et al. 2016). The crucial importance of HDR is in deploying effective techniques to detect the written digits in a more accurate and speedy manner. HDR is also used as a benchmark for comparing different classification techniques. Specifically, the optical character recognition (ORC) is a sub-task of HDR problem that includes some applications such as postal code and bank cheque reading, automatic pin code recognition and collecting data from filling in forms (Maheshwari et al. 2018). The Communicated by V. Loia.
B 1
Rami S. Alkhawaldeh [email protected] Department of Computer Information Systems, The University of Jordan, Aqaba 77110, Jordan
original works of OCR handwritten alphabets and numerals started with Latin languages. Several works show an efficient performance related to these languages using printed characters rather than handwritten ones. Although these works achieved better performance, other languages such as Arabic are in the exploration and analysis pha
Data Loading...
