OCR-Nets: Variants of Pre-trained CNN for Urdu Handwritten Character Recognition via Transfer Learning

K.O, Mohammed Aarif; Poruran, Sivakumar

doi:10.1016/j.procs.2020.04.248

Cited by 36 publications

(11 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The quantitative results for hybrid dataset are presented in [19] 92.09% 7.1% Neural network [19] 91.98% 8.02% SVM [19] 95.79% 4.21% CNN [19] 96.7% 3.3% Autoencoder [19] 97.3% 2.7% Guassian NB [19] 69.3% 30.7% Decision Tree [19] 82.00% 18% Daubechies wavelet [59] 92.05% 7.95% Fuzzy rule [60] 97.4% 2.6% HMM (Hidden Markov Model) [60] 96.2% 3.8% Hybrid approach (Fuzzy and HMM) [60] 97.8% 2.5% Fuzzy rule base approach [61] 96.3% [19] 81.2% ---18.8% CNN [19] 86.6% ---13.4% CNN [33] 80.7% ---19.3% SVM [33] 56% ---44.0% ANN [33] 78% ---22.0% LeNet [62] 90.34% ---9.70% SVM with polynomial kernel on 40x36 [63] 88.80 % ---11.20% SVM with Transfer Learning [64] 82.30% ---17.7% BLSTM [65] 92-94% ---8 -6% OCR-GoogleNet [64] 94.7% to be effective in a variety of domains of recognition tasks, computer vision, including textile image analysis, in recent years.…”

Section: Results For Hybrid Datasetmentioning

confidence: 99%

Handwritten Urdu Characters and Digits Recognition Using Transfer Learning and Augmentation With AlexNet

et al. 2022

View full text Add to dashboard Cite

Automated recognition of handwritten characters and digits is a challenging task. Although a significant amount of literature exists for automatic recognition of handwritten characters of English and other major languages in the world, there exists a wide research gap due to lack of research for recognition of Urdu language. The variations in writing style, shape and size of individual characters and similarities with other characters add to the complexity for accurate classification of handwritten characters. Deep neural networks have emerged as a powerful technology for automated classification of character patters and object images. Although deep networks are known to provide remarkable results on large-scale datasets with millions of images, however the use of deep networks for small image datasets is still challenging. The purpose of this research is to present a classification framework for automatic recognition of handwritten Urdu character and digits with higher recognition accuracy by utilizing theory of transfer learning and pretrained Convolution Neural Networks (CNN). The performance of transfer learning is evaluated in different ways: by using pre-trained AlexNet CNN model with Support Vector Machine (SVM) classifier, and finetuned AlexNet for extracting features and classification. We have fine-tuned AlexNet hyper-parameters to achieve higher accuracy and data augmentation is performed to avoid over-fitting. Experimental results and the quantitative comparisons demonstrate the effectiveness of the proposed research for recognition of handwritten characters and digits using fine-tuned AlexNet. The proposed research based on fine-tuned AlexNet outperforms the related state-of-the-art research thereby achieving a classification accuracy of 97.08%, 98.21%, 94.92% for urdu characters, digits and hybrid datasets respectively. The presented methods can be applied for research on Urdu characters and in diverse domains such as handwritten text image retrieval, reading postal addresses, bank's cheque processing, preserving and digitization of manuscripts from old ages.

show abstract

Section: Results For Hybrid Datasetmentioning

confidence: 99%

Handwritten Urdu Characters and Digits Recognition Using Transfer Learning and Augmentation With AlexNet

et al. 2022

View full text Add to dashboard Cite

show abstract

“…A third solution in this context lies in the use of TL [64][65][66]76]. TL was less widely adopted in AHR compared to other techniques, and its benefits have been limited due to the use of models pre-trained on ImageNet, a dataset distant from the specifics of AHR.…”

Section: Highlights and Discussionmentioning

confidence: 99%

“…Arif and Poruran [66] employed TL in two architectures referred to as OCR-AlexNet and OCR-GoogleNet, adapted from the original architectures of AlexNet and GoogleNet. In the case of OCR-AlexNet, the initial layer weights of AlexNet were retained, while adjustments were made to the final three layers to tailor them to the new task.…”

Section: Transfer Learningmentioning

confidence: 99%

CNN-based Methods for Offline Arabic Handwriting Recognition: A Review

El Khayati,

Kich,

Taouil

2024

Neural Process Lett

View full text Add to dashboard Cite

Arabic Handwriting Recognition (AHR) is a complex task involving the transformation of handwritten Arabic text from image format into machine-readable data, holding immense potential across various applications. Despite its significance, AHR encounters formidable challenges due to the intricate nature of Arabic script and the diverse array of handwriting styles. In recent years, Convolutional Neural Networks (CNNs) have emerged as a pivotal and promising solution to address these challenges, demonstrating remarkable performance and offering distinct advantages. However, the dominance of CNNs in AHR lacks a dedicated comprehensive review in the existing literature. This review article aims to bridge the existing gap by providing a comprehensive analysis of CNN-based methods in AHR. It covers both segmentation and recognition tasks, delving into advancements in network architectures, databases, training strategies, and employed methods. The article offers an in-depth comparison of these methods, considering their respective strengths and limitations. The findings of this review not only contribute to the current understanding of CNN applications in AHR but also pave the way for future research directions and improved practices, thereby enriching and advancing this critical domain. The review also aims to uncover genuine challenges in the domain, providing valuable insights for researchers and practitioners.

show abstract

“…In this matrix, each pixel is an element of the matrix. In a three channel color image, this element is a tuple of three numbers [6][7][8][9].…”

Section: Operating Principle and Characteristics Of Ocr Systemmentioning

confidence: 99%

Electronic Bill Fast Recognition Based on OCR and CNN

Zheng,

Chen,

et al. 2024

Advances in Transdisciplinary Engineering

View full text Add to dashboard Cite

The existing methods used for bill recognition need specific equipment scanning or a large number of bill labels to achieve a good recognition effect. To solve above problems, a bill recognition scheme based on OCR technology is proposed. In the part of image preprocessing, the methods of binarization, filtering and denoising, image rectification are used. Then, we use deep learning as the basic model of the bill text detection and recognition phase, and use ResNet to complete the detection of specific fields in the way of image classification. Finally, on the basis of the rapid identification method designed in this paper, the feasibility and effectiveness of the scheme are verified through the demand analysis and functional test of the bill identification system. The results show that it has achieved good results in both simulation data sets and practical applications, and it has more advantages than traditional recognition technologies in text and image recognition.

show abstract

OCR-Nets: Variants of Pre-trained CNN for Urdu Handwritten Character Recognition via Transfer Learning

Cited by 36 publications

References 11 publications

Handwritten Urdu Characters and Digits Recognition Using Transfer Learning and Augmentation With AlexNet

Handwritten Urdu Characters and Digits Recognition Using Transfer Learning and Augmentation With AlexNet

CNN-based Methods for Offline Arabic Handwriting Recognition: A Review

Electronic Bill Fast Recognition Based on OCR and CNN

Contact Info

Product

Resources

About