检索结果-内蒙古大学图书馆

Tamil Language Handwritten document digitization and Analysis of the Impact of Data Augmentation Using Generative Adversarial Networks (GANs) on the Accuracy of CNN Model 1

引用

International Conference on Machine Learning and Autonomous Systems (ICMLAS)

作者： Murugesh, Venkatesh Parthasarathy, Aditya Gopinath, Gokul P. Khade, Anindita SIES Grad Sch Technol Dept Comp Engn Navi Mumbai India

ISBN: (数字)9789811679964

ISBN: (纸本)9789811679964;9789811679957

Recognition of handwritten documents is an essential part of today'sworld for various reasons such as analysis, data extraction, etc. Recognition of handwritten in Indian regional languages, for, e.g., Tamil character is very difficult, due to variations in size, style, and orientation angle. Limited availability of handwritten character dataset makes it difficult to achieve high accuracy for all the characters present in the language. In this paper, we propose a deep learning approach to recognize and classify handwritten characters of the Tamil language. The proposed system can be divided into five phases: expansion of the dataset and data augmentation with the help of Generative Adversarial Networks (GAN), preprocessing of the input image to reduce the noise, segmentation of the characters, feature extraction, and classification of the character done using Convolutional Neural Networks (CNN). We have achieved an accuracy of 94.03% and with data augmentation using GAN, an accuracy of 97% was achieved. The proposed system, along with the recognition of characters, also converts the image of the handwritten document into an editable word document or a PDF file.

关键词： Character segmentation CNN Deep learning document digitization GAN Generative adversarial network Tamil character recognition

来源：评论

学校读者我要写书评

暂无评论

Robust Camera Captured Image Mosaicking for document digitization and OCR Processing 13

Robust Camera Captured Image Mosaicking for Document Digitiz...

引用

International Conference on Information Technology (ICIT)

作者： Tiwari, Lokender Kumar, Bhupendra Patnaik, Tushar Ctr Dev Adv Comp Sch IT Noida 201307 India

ISBN: (纸本)9781479980840

People often capture document images from their mobile camera or handheld digital camera. These cameras are handy and can capture partial images of a big document like chart or new paper and text written on wall. Image mosaicking is the process of reconstructing the whole image from its constituent partial overlapped images. If partial images of a document are captured by flatbed scanner then mosaic of the document is obtained by simply estimating the homography which is a simple 2D transformation between partial images. While this is not the case when partial images are captured by mobile or handheld camera. Camera images are suffered from illumination change, scale change, change in viewpoint, perspective distortions and low resolution. In this paper we discussed a robust mosaicking approach for camera captured document images and its implementation results for the applications like document digitization and OCR processing.

关键词： document digitization mosaic camera captured image mosaicking robust homography affine correction projective correction

来源：评论

学校读者我要写书评

暂无评论

OCR-CRNN (WBS): an optical character recognition system based on convolutional recurrent neural network embedded with word beam search decoder for extraction of text

引用

International Journal of Information Technology (Singapore) 2025年 1-8页

作者： Kumar, Mahesh Shalu Dureja, Ajay Narula, Rachna Shyla Arora, Ravi Department of Information Technology Bharati Vidyapeeth’s College of Engineering (BVCOE) New Delhi India Department of Computer Science and Technology Manav Rachna University Haryana Faridabad India Department of Computer Science & Engineering Bharati Vidyapeeth’s College of Engineering New Delhi India

With the growing need for accurate digital text retrieval from images, OCR (optical character recognition) plays a critical role in fields such as document digitization, automated data processing, and information retrieval. The given paper introduces optical OCR system, comprising Convolutional Recurrent Neural Network (CRNN) embedded with a Word Beam Search (WBS) decoder for extraction of text. The framework performs preprocessing, feature extraction, and post-processing strategies that enhance the accuracy and robustness of text extraction. Preprocessing techniques and post-processing adjustments significantly improved the model's ability to handle diverse image formats, including scanned documents and screenshots. The proposed system demonstrates an overall accuracy of 92%, with macro and weighted averages across precision, recall, and F1-scores also at 92%. Individual class evaluations reveal notable performance consistency, with precision ranging from 0.79 to 0.98, recall between 0.79 and 0.96, and F1-scores from 0.79 to 0.94 across 26 alphabetic characters. This research highlights the system's potential for real-world applications, demonstrating its scalability, precision, and efficiency in extracting text data across various formats, ensuring high OCR accuracy in automated data processing. © Bharati Vidyapeeth's Institute of Computer Applications and Management 2025.

关键词： CRNN document digitization Optical character recognition (OCR) Text extraction WBS

来源：评论

学校读者我要写书评

暂无评论

A Hybrid Approach for Table Information Extraction: Combining Google document AI with Custom Object Detection Models 10

A Hybrid Approach for Table Information Extraction: Combinin...

引用

10th IEEE International Conference on Electronics, Computing and Communication Technologies (IEEE CONECCT)

作者： Balakrishnan, Muthukumaraswamy Rajagopal, Ranjithkumar Searce Appl AI Bengaluru India

ISBN: (纸本)9798350385939;9798350385922

In today's era of document digitization, the extraction of table data from both structured and unstructured tables documents holds significant importance. Tables found within documents such as scientific articles, invoices, and others often exhibit diverse formats and designs, varying from those with borders to those without, or with minimal borders. Addressing this variability necessitates the development of a custom mechanism for extracting data from these tables. This paper proposes a tailored solution to address such extraction challenges. The proposed approach involves the collection of custom data specifically tailored for extraction purposes. Subsequently, three object detection models (Yolov5) are trained to identify regions of interest (ROIs), annotate rows and columns within tables, and train with the collected custom data. Pretrained models are used to ease the job of annotation. Through inference and post-processing of the model results, cell bounding boxes are generated. Leveraging these bounding boxes, we can align the output of Google document AI with the specific document, facilitating accurate extraction of desired results. This custom solution offers a robust method for efficiently extracting table data from documents with diverse formats and designs. The results and performance evaluation proves that the method proposed is competent in real time.

关键词： document digitization Table data extraction Object detection Custom solution document AI

来源：评论

学校读者我要写书评

暂无评论

Figure search by text in large scale digital document collections

引用

CONCURRENCY AND COMPUTATION-PRACTICE & EXPERIENCE 2022年第1期34卷

作者： Yurtsever, M. Mucahit Enes Ozcan, Muhammet Taruz, Zubeyir Eken, Suleyman Sayar, Ahmet Kocaeli Univ Dept Informat Syst Engn Umuttepe Campus TR-41001 Kocaeli Turkey Kocaeli Univ Dept Comp Engn Kocaeli Turkey

Digital document collections have been created with the transfer of a large number of documents to digital media. These digital archives have provided many benefits to users. As the diversity and size of digital image collections have grown exponentially, it has become increasingly important and difficult to obtain the desired image from them. The images on the document might contain critical information about the subject of it. In this study, an architecture is developed that can work on large-scale data by creating regular expressions together with full-text search approaches. The performance of the system has been tested on different academic documents and Elasticsearch and Apache Solr insert times are compared. Compared to Elasticsearch, Apache Solr achieved faster and more successful results.

关键词： Apache Solr document digitization Elasticsearch figure search full-text search regular expressions

来源：评论

学校读者我要写书评

暂无评论

A New Method for Detecting Altered Text in document Images

引用

INTERNATIONAL JOURNAL OF PATTERN RECOGNITION AND ARTIFICIAL INTELLIGENCE 2021年第12期35卷

作者： Nandanwar, Lokesh Shivakumara, Palaiahnakote Pal, Umapada Lu, Tong Lopresti, Daniel Seraogi, Bhagesh Chaudhuri, Bidyut B. Univ Malaya Fac Comp Sci & Informat Technol Kuala Lumpur Malaysia Indian Stat Inst Comp Vis & Pattern Recognit Unit Kolkata India Nanjing Univ Natl Key Lab Novel Software Technol Nanjing Peoples R China Lehigh Univ Comp Sci & Engn Bethlehem PA 18015 USA Techno India Univ Comp Sci & Engn Sect 5 Kolkata India

As more and more office documents are captured, stored, and shared in digital format, and as image editing software are becoming increasingly more powerful, there is a growing concern about document authenticity. To prevent illicit activities, this paper presents a new method for detecting altered text in document images. The proposed method explores the relationship between positive and negative coefficients of DCT to extract the effect of distortions caused by tampering by fusing reconstructed images of respective positive and negative coefficients, which results in Positive-Negative DCT coefficients Fusion (PNDF). To take advantage of spatial information, we propose to fuse R, G, and B color channels of input images, which results in RGBF (RGB Fusion). Next, the same fusion operation is used for fusing PNDF and RGBF, which results in a fused image for the original input one. We compute a histogram to extract features from the fused image, which results in a feature vector. The feature vector is then fed to a deep neural network for classifying altered text images. The proposed method is tested on our own dataset and the standard datasets from the ICPR 2018 Fraud Contest, Altered Handwriting (AH), and faked IMEI number images. The results show that the proposed method is effective and the proposed method outperforms the existing methods irrespective of image type.

关键词： document digitization DCT coefficients fused image altered text detection fraud document CNN

来源：评论

学校读者我要写书评

暂无评论

Data Collection and Image Processing System for Ancient Arabic Manuscripts 2

Data Collection and Image Processing System for Ancient Arab...

引用

2nd IEEE International Workshop on Arabic and Derived Script Analysis and Recognition (IEEE ASAR)

作者： Al-Maadeed, Somaya Peer, Syed F. K. Subramanian, Nandhini Qatar Univ Dept Comp Sci & Engn Doha Qatar

ISBN: (纸本)9781538614594

This paper presents a general-purpose data collection system that combines a DSLR camera with directional LED lamps in order to capture a large quantity of high-resolution manuscript images in such a way as to maximize the speed of data collection while minimizing time and the need for specialized equipment. By integrating custom image processing software, the captured document images are mapped to lie on a planar surface, thereby enabling the application of more sophisticated computer vision algorithms. For this purpose, we also introduce an optional binarization tool that allows researchers to perform basic image pre-processing to simplify later analysis. The hardware setup and software tools presented in this paper can be combined to yield a simple system capable of producing large image datasets for use in document analysis research projects.

关键词： document digitization hardware design image post-processing image pre-processing scanning software

来源：评论

学校读者我要写书评

暂无评论

Non-Destructive digitization of Soiled Historical Chinese Bamboo Scrolls 13

Non-Destructive Digitization of Soiled Historical Chinese Ba...

引用

13th IAPR International Workshop on document Analysis Systems (DAS)

作者： Stromer, Daniel Christlein, Vincent Maier, Andreas Zippert, Patrick Helmecke, Eric Hausotte, Tino Huang, Xiaolin Univ Erlangen Nuremberg FAU Pattern Recognit Lab Erlangen Germany Univ Erlangen Nuremberg FAU Inst Mfg Metrol Erlangen Germany Shanghai Jiao Tong Univ Inst Image Proc & Pattern Recognit Shanghai Peoples R China

ISBN: (纸本)9781538633465

For about 2000 years, no paper was used as a media in China but writings and drawings were captured on bamboo and wooden slips. Several slips were bound together with strips and rolled up to a scroll. The writings and drawings were either brushed or even carved into the wood. Those documents are very precious for culture inheritance and research, but due to aging processes, the discovered pieces are sometimes in a poor condition and also soiled. Because cleaning the slips is not only challenging but also writings could be erased, we developed a method to digitize such historical documents without the need of cleaning. We perform a 3-D X-ray micro-CT scan resulting in a 3-D volume of the complete document. With our approach, we were able to investigate the scroll without any manual labor (e.g. unwrapping or cleaning). We showed that the method also works for heavily soiled scrolls where nothing is readable with the naked eye. This can help conservators to store all writings before they may be erased by the cleaning process. Finally, we present a manual technique to virtually unwrap and post-process the documents resulting in a 2-D image of all bamboo slips.

关键词： Bamboo and wooden slips 3-D X-ray CT Historical Chinese documents document digitization

来源：评论

学校读者我要写书评

暂无评论

Separating Tables from Text and non-Text Objects in Printed documents for Digital Reconstruction 12

Separating Tables from Text and non-Text Objects in Printed ...

引用

IEEE International Conference on Industrial and Information Systems (ICIIS)

作者： Jahan, M. A. C. Akmal Ragel, R. G. Univ Peradeniya Postgrad Inst Sci Peradeniya Sri Lanka Univ Peradeniya Dept Comp Engn Peradeniya Sri Lanka

ISBN: (纸本)9781538616765

Reproducing printed documents with their existing format is becoming an important task when we need to reprint or republish existing printed documents with the latest updates of contents. Text, images, charts, graphs, tables, logos and signatures are some of the prominent components of a printed document. When reproducing the text of a printed document, optical character recognition techniques can address the text related issues such as detection, recognition, and reconstruction. Beyond that, tables can be made editable if it is detected correctly, recognized and positioned in the document. Researchers who have addressed table-related issues have only dealt with tables when they are in line with the text. There is rare work in locating tables when a printed material has text, tables, and other non-text components all in the same document. A few available works treat tables as non-text objects and, therefore, no mechanisms were presented to separate them from the non-text elements. Considering tables as non-text component make it impossible to edit and update them in the usual way how tables are edited and updated. Therefore, our work addresses this issue by separating tables, text and other non-text objects within a document image using rule based thresholds and then by reconstructing the tables with the extracted features. From the experiment carried out on about 480 document images, we have achieved 81% of automated table detection accuracy when a printed material has text, table, and non-text components. Moreover, we have achieved about 90% accuracy when we input the type information of the document manually.

关键词： table detection table reconstruction document digitization

来源：评论

学校读者我要写书评

暂无评论

An Intranet-Based document Management and Monitoring System Framework: A Case for the National University Quality Management Office

An Intranet-Based Document Management and Monitoring System ...

引用

IEEE Region 10 Conference (TENCON)

作者： Minon, Judi Diane F. Lim, Christine Mae A. Morano, Julie Ann L. Fajutagana, Raymart F. Fabito, Bernie S. Natl Univ Coll Comp Studies Manila Philippines

ISBN: (纸本)9781509025978

Universities typically hold large amounts of documents that contain important and sensitive data. These documents are needed to be kept securely as it passes from one office to another, restricting any unauthorized access. Manual submission of papers to offices revealed that it is subjected to being unmonitored, misplaced and takes a lot of time to be processed if not properly supervised. This paper focuses on the need for the design and development of an intranet based website that creates, transmits, monitors, and restricts ISO documents that go in and out of National University. Old physical documents will be scanned to convert into digital copies and will be uploaded to the system and will go to its proper repository depending on its classification based on the file name convention. A digital signature is also included for signatories which can be tracked and monitored. Additionally, it provides a framework where the development of an Intranet based documents Management System can be based.

关键词： Records Management document digitization Digital Signature

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：