A Novel Deep Convolutional Neural Network Architecture Based on Transfer Learning for Handwritten Urdu Character Recognition

Deep convolutional neural networks (CNN) have made a huge impact on computer vision and set the state-of-the-art in providing extremely definite classification results. For character recognition, where the training images are usually inadequate, mostly transfer learning of pre-trained CNN is often u...

Full description

Saved in:
Bibliographic Details
Published inTehnički vjesnik Vol. 27; no. 4; pp. 1160 - 1165
Main Authors Poruran, Sivakumar, Caffiyar, Mohamed Yousuff
Format Journal Article Paper
LanguageEnglish
Published Slavonski Baod Josipa Jurja Strossmayer University of Osijek 15.08.2020
Strojarski fakultet u Slavonskom Brodu; Fakultet elektrotehnike, računarstva i informacijskih tehnologija Osijek; Građevinski i arhitektonski fakultet Osijek
Faculty of Mechanical Engineering in Slavonski Brod, Faculty of Electrical Engineering in Osijek, Faculty of Civil Engineering in Osijek
Subjects
Artificial neural networks
Character recognition
Computer architecture
Computer vision
deep learning
document analysis
Extreme values
Feature extraction
Handwriting recognition
Image classification
Learning
Neural networks
Object recognition
pattern recognition
pretrained CNN
transfer learning
Urdu character recognition
Online AccessGet full text

Cover

More Information
Summary:Deep convolutional neural networks (CNN) have made a huge impact on computer vision and set the state-of-the-art in providing extremely definite classification results. For character recognition, where the training images are usually inadequate, mostly transfer learning of pre-trained CNN is often utilized. In this paper, we propose a novel deep convolutional neural network for handwritten Urdu character recognition by transfer learning three pre-trained CNN models. We fine-tuned the layers of these pre-trained CNNs so as to extract features considering both global and local details of the Urdu character structure. The extracted features from the three CNN models are concatenated to train with two fully connected layers for classification. The experiment is conducted on UNHD, EMILLE, DBAHCL, and CDB/Farsi dataset, and we achieve 97.18% average recognition accuracy which outperforms the individual CNNs and numerous conventional classification methods.
Bibliography:242316
ISSN:1330-3651
1848-6339
DOI:10.17559/TV-20190319095323