Analysis of Recent Deep Learning Techniques for Arabic Handwritten-Text OCR and Post-OCR Correction

Najam, Rayyan Qari Shahabuddin; Faizullah, Safiullah

doi:10.3390/app13137568

Cited by 9 publications

(4 citation statements)

References 75 publications

(107 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Faizullah et al [22] presented a review of Arabic recognition, where they presented the entire process of optical character recognition (OCR) for Arabic, including some of the most advanced techniques, and provided future research directions for Arabic recognition. After this, Najam et al [23] provided the latest deep learning techniques for OCR technology and correction after OCR for Arabic handwriting, developing a practical space for future trends in Arabic OCR applications.…”

Section: Related Workmentioning

confidence: 99%

A Three-Stage Uyghur Recognition Model Combining the Attention Mechanism and Different Convolutional Recurrent Networks

Li,

Zhang,

Huang

et al. 2023

Applied Sciences

View full text Add to dashboard Cite

Uyghur text recognition faces several challenges in the field due to the scarcity of publicly available datasets and the intricate nature of the script characterized by strong ligatures and unique attributes. In this study, we propose a unified three-stage model for Uyghur language recognition. The model is developed using a self-constructed Uyghur text dataset, enabling evaluation of previous Uyghur text recognition modules as well as exploration of novel module combinations previously unapplied to Uyghur text recognition, including Convolutional Recurrent Neural Networks (CRNNs), Gated Recurrent Convolutional Neural Networks (GRCNNs), ConvNeXt, and attention mechanisms. Through a comprehensive analysis of the accuracy, time, normalized edit distance, and memory requirements of different module combinations on a consistent training and evaluation dataset, we identify the most suitable text recognition structure for Uyghur text. Subsequently, utilizing the proposed approach, we train the model weights and achieve optimal recognition of Uyghur text using the ConvNeXt+Bidirectional LSTM+attention mechanism structure, achieving a notable accuracy of 90.21%. These findings demonstrate the strong generalization and high precision exhibited by Uyghur text recognition based on the proposed model, thus establishing its potential practical applications in Uyghur text recognition.

show abstract

Section: Related Workmentioning

confidence: 99%

A Three-Stage Uyghur Recognition Model Combining the Attention Mechanism and Different Convolutional Recurrent Networks

Li,

Zhang,

Huang

et al. 2023

Applied Sciences

View full text Add to dashboard Cite

show abstract

“…Deep learning strategies have reformed the field of optical character recognition (OCR) and enhanced the accuracy in image-to-LaTeX conversion. Leveraging neural network architectures like recurrent neural networks and convolutional neural networks enables researchers to tackle the complex challenges of character recognition and equation parsing [13,14]. The learning models rely on large-scale annotated datasets to learn and generalize.…”

Section: Deep Learning Strategies For Ocr In Image-to-latex Conversionmentioning

confidence: 99%

Advancing OCR Accuracy in Image-to-LaTeX Conversion—A Critical and Creative Exploration

Orji,

Haydar,

Erşan

et al. 2023

Applied Sciences

View full text Add to dashboard Cite

This paper comprehensively assesses the application of active learning strategies to enhance natural language processing-based optical character recognition (OCR) models for image-to-LaTeX conversion. It addresses the existing limitations of OCR models and proposes innovative practices to strengthen their accuracy. Key components of this study include the augmentation of training data with LaTeX syntax constraints, the integration of active learning strategies, and the employment of active learning feedback loops. This paper first examines the current weaknesses of OCR models with a particular focus on symbol recognition, complex equation handling, and noise moderation. These limitations serve as a framework against which the subsequent research methodologies are assessed. Augmenting the training data with LaTeX syntax constraints is a crucial strategy for improving model precision. Incorporating symbol relationships, wherein contextual information is considered during recognition, further enriches the error correction. This paper critically examines the application of active learning strategies. The active learning feedback loop leads to progressive improvements in accuracy. This article underlines the importance of uncertainty and diversity sampling in sample selection, ensuring that the dynamic learning process remains efficient and effective. Appropriate evaluation metrics and ensemble techniques are used to improve the operational learning effectiveness of the OCR model. These techniques allow the model to adapt and perform more effectively in diverse application domains, further extending its utility.

show abstract

“…To study manuscripts affected by ink loss or brown spotting, multispectral imaging techniques such UV fluorescence and VIS-NIR imaging can be employed [3,4]. Text enhancement and segmentation may leverage deep learning [5,6]. We present a method for handling images from digital libraries, which typically consist of RGB images captured by portable digital cameras or flatbed scanners [7].…”

Section: Introductionmentioning

confidence: 99%

Assessing readability of the text in ancient paper fragments by a photometric statistical analysis

Franchi,

Colonnese,

Cedola

et al. 2024

J. Inst.

View full text Add to dashboard Cite

Ancient documents are important historical sources that are often found in a fragmented condition due to their conservation status. In this study, we examined fragments of paper found in 1996 during excavation of the Santi Quattro Coronati complex, in Rome. The archaeological site where the fragments were found is situated on the first floor of the tower within the complex. This location was used as a disposal pit approximately between the 15th and 16th centuries. The fragments exhibit text discoloration, hindering automatic recognition and human readability. To reveal the faded text, the fragments have been digitalized, converted into a perceptually uniform color space and the contrast has been enhanced. The photometric characteristics of the input and enhanced images have been statistically characterized, and the contrast enhancement assessed by a state-of-the-art metric. The statistical analysis of the text colour coordinates was carried out to develop supervised and unsupervised image segmentation, isolating the text. The results of the method show that it effectively identifies text regions within images, improving readability, even for faded text. It can be integrated into deep learning-based character recognition systems, facilitating the automatic analysis of historical handwritten documents.

show abstract

Analysis of Recent Deep Learning Techniques for Arabic Handwritten-Text OCR and Post-OCR Correction

Cited by 9 publications

References 75 publications

A Three-Stage Uyghur Recognition Model Combining the Attention Mechanism and Different Convolutional Recurrent Networks

A Three-Stage Uyghur Recognition Model Combining the Attention Mechanism and Different Convolutional Recurrent Networks

Advancing OCR Accuracy in Image-to-LaTeX Conversion—A Critical and Creative Exploration

Assessing readability of the text in ancient paper fragments by a photometric statistical analysis

Contact Info

Product

Resources

About