Segmentation of Scanned Documents Using Deep-Learning Approach

Forczmański, Paweł; Smoliński, Anton; Nowosielski, Adam; Małecki, Krzysztof

doi:10.1007/978-3-030-19738-4_15

Cited by 10 publications

(5 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Some of these criteria are also outlined in the WCAG as well. Research in this field focuses on identifying potential layout errors and using various automatic methods to reduce them (Shafait, 2008; Erkilinc et al , 2011; Forczmański et al , 2020; Xu et al , 2020; Zhong et al , 2019).…”

Section: Related Researchmentioning

confidence: 99%

Improving accessibility of digitization outputs: EODOPEN project research findings

Kavčič Čolić,

Hari

2024

DLP

View full text Add to dashboard Cite

Purpose The current predominant delivery format resulting from digitization is PDF, which is not appropriate for the blind, partially sighted and people who read on mobile devices. To meet the needs of both communities, as well as broader ones, alternative file formats are required. With the findings of the eBooks-On-Demand-Network Opening Publications for European Netizens project research, this study aims to improve access to digitized content for these communities. Design/methodology/approach In 2022, the authors conducted research on the digitization experiences of 13 EODOPEN partners at their organizations. The authors distributed the same sample of scans in English with different characteristics, and in accordance with Web content accessibility guidelines, the authors created 24 criteria to analyze their digitization workflows, output formats and optical character recognition (OCR) quality. Findings In this contribution, the authors present the results of a trial implementation among EODOPEN partners regarding their digitization workflows, used delivery file formats and the resulting quality of OCR results, depending on the type of digitization output file format. It was shown that partners using the OCR tool ABBYY FineReader Professional and producing scanning outputs in tagged PDF and PDF/UA formats achieved better results according to set criteria. Research limitations/implications The trial implementations were limited to 13 project partners’ organizations only. Originality/value This research paper can be a valuable contribution to the field of massive digitization practices, particularly in terms of improving the accessibility of the output delivery file formats.

show abstract

Section: Related Researchmentioning

confidence: 99%

Improving accessibility of digitization outputs: EODOPEN project research findings

Kavčič Čolić,

Hari

2024

DLP

View full text Add to dashboard Cite

show abstract

“…[13] use a conditional random field model for the prediction and achieve an average accuracy of 0.83 for all 13 classes which they distinguish. Scientific works, in which Convolutional Neural Networks (CCNs) are used for the classification task of image-based document pages, are reported by [14], [15], and [16], among others. [14] proposed a method (document domain randomization (DDR)) that does not need manually annotated document pages, but works with generated pseudo-pages.…”

Section: Related Workmentioning

confidence: 99%

“…"email", "news article", "file folder", "letters", "memo", and so on. [16] performs object detection on image-based document pages using CNN. The objects they want to identify are "stamps", "logos", "signatures", "tables" and "text blocks".…”

Section: Related Workmentioning

confidence: 99%

Preparing Legal Documents for NLP Analysis: Improving the Classification of Text Elements by Using Page Features

Josi¹,

Wartena²,

Heid³

2022

Natural Language Processing (NLP) Trends

View full text Add to dashboard Cite

Legal documents often have a complex layout with many different headings, headers and footers, side notes, etc. For the further processing, it is important to extract these individual components correctly from a legally binding document, for example a signed PDF. A common approach to do so is to classify each (text) region of a page using its geometric and textual features. This approach works well, when the training and test data have a similar structure and when the documents of a collection to be analyzed have a rather uniform layout. We show that the use of global page properties can improve the accuracy of text element classification: we first classify each page into one of three layout types. After that, we can train a classifier for each of the three page types and thereby improve the accuracy on a manually annotated collection of 70 legal documents consisting of 20,938 text elements. When we split by page type, we achieve an improvement from 0.95 to 0.98 for single-column pages with left marginalia and from 0.95 to 0.96 for double-column pages. We developed our own feature-based method for page layout detection, which we benchmark against a standard implementation of a CNN image classifier.

show abstract

“…Many recent works are available for the proposed problem, which use neural networks or deep learning models. For example, Forczmański et al [8] presented an object detection approach using a Convolutional Neural Network. They focused on automatic segmentation of elements from documents.…”

Section: Related Work and Motivationmentioning

confidence: 99%

Robust Detection of Tables in Documents Using Scores from Table Cell Cores

et al. 2022

View full text Add to dashboard Cite

Table detection is an essential step in many document analysis systems. Tabular data are a pivotal form of information representation that can organize data in a conventional structure for comfortable and quick information retrieval and comparison. Detection of table structures in PDF files or images is a challenging task because of the variability of table layouts, and sometimes the tabular structures’ similarities with non-tabular elements like charts, plots, etc. In this work, we have presented a table detection method using a geometric analysis of the table cell cores that represents the table cell texts. The proposed method works by analyzing the text gap information, and hence it can detect the table cell cores, irrespective of the presence of the table boundary lines and cell-separating rule-lines. Experimentations have been done on various document images of complex structures from well-known datasets. The detection accuracies obtained by us corroborate the usefulness of the proposed method.

show abstract

Segmentation of Scanned Documents Using Deep-Learning Approach

Cited by 10 publications

References 18 publications

Improving accessibility of digitization outputs: EODOPEN project research findings

Improving accessibility of digitization outputs: EODOPEN project research findings

Preparing Legal Documents for NLP Analysis: Improving the Classification of Text Elements by Using Page Features

Robust Detection of Tables in Documents Using Scores from Table Cell Cores

Contact Info

Product

Resources

About