Filters








589 Hits in 4.1 sec

Recognizing Challenging Handwritten Annotations with Fully Convolutional Networks [article]

Andreas Kölsch, Ashutosh Mishra, Saurabh Varshneya, Muhammad Zeshan Afzal, Marcus Liwicki
2018 arXiv   pre-print
This paper introduces a very challenging dataset of historic German documents and evaluates Fully Convolutional Neural Network (FCNN) based methods to locate handwritten annotations of any kind in these  ...  For evaluation, we use the Layout Analysis Evaluator for the ICDAR 2017 Competition on Layout Analysis for Challenging Medieval Manuscripts.  ...  Finally, we thank the following libraries for providing us with the digitized images for the database: Universitätsbibliothek der Humboldt-Universität zu Berlin; Universitätsbibliothek Kassel; Staatsbibliothek  ... 
arXiv:1804.00236v2 fatcat:yyqrfrld7nacjkq4cbl4utwtry

Historical Handwritten Document Segmentation by Using a Weighted Loss [chapter]

Samuele Capobianco, Leonardo Scommegna, Simone Marinai
2018 Lecture Notes in Computer Science  
In this work we propose one deep architecture to identify text and not-text regions in historical handwritten documents.  ...  We obtain good results using global metrics improving global and local classification scores.  ...  A Fully Convolution Network has been also used for page segmentation [14] where the FCN is used to provide a pixel-wise classification followed by post processing techniques to split a document image  ... 
doi:10.1007/978-3-319-99978-4_31 fatcat:p3m7xrjszffkxctznbkuk7oirq

Text Line Segmentation for Challenging Handwritten Document Images Using Fully Convolutional Network [article]

Berat Barakat, Ahmad Droby, Majeed Kassis, Jihad El-Sana
2021 arXiv   pre-print
Then these line masks are predicted using a Fully Convolutional Network (FCN). In the literature, FCN has been successfully used for text line segmentation of regular handwritten document images.  ...  This paper presents a method for text line segmentation of challenging historical manuscript images.  ...  ACKNOWLEDGMENT The authors would like to thank the support of the Frankel Center for Computer Science at Ben-Gurion University of the Negev.  ... 
arXiv:2101.08299v1 fatcat:bmzohu5mzrfititec7z2vlcdum

Deep Learning for Historical Document Analysis and Recognition—A Survey

Francesco Lombardi, Simone Marinai
2020 Journal of Imaging  
The analysis and recognition of historical documents, as we survey in this work, is not an exception.  ...  Our study analyzes the papers published in the last few years on this topic from different perspectives: we first provide a pragmatic definition of historical documents from the point of view of the research  ...  Fully Convolutional Networks The basic concept behind Fully Convolutional Networks (FCN) is to contain only convolutional layers [86] .  ... 
doi:10.3390/jimaging6100110 pmid:34460551 pmcid:PMC8321201 fatcat:nevh2ctshzfwtey4girgjtaftq

Convolutional Neural Networks for Page Segmentation of Historical Document Images [article]

Kai Chen, Mathias Seuret
2017 arXiv   pre-print
This paper presents a Convolutional Neural Network (CNN) based page segmentation method for handwritten historical document images.  ...  We consider page segmentation as a pixel labeling problem, i.e., each pixel is classified as one of the predefined classes.  ...  CONCLUSION In this paper, we have proposed a convolutional neural network (CNN) for page segmentation of handwritten historical document images.  ... 
arXiv:1704.01474v2 fatcat:dvmojuaknvc2pignzzo3bznif4

In Codice Ratio: OCR of Handwritten Latin Documents using Deep Convolutional Networks

Donatella Firmani, Paolo Merialdo, Elena Nieddu, Simone Scardapane
2017 International Conference of the Italian Association for Artificial Intelligence  
Automatic transcription of historical handwritten documents is a challenging research problem, requiring in general expensive transcriptions from expert paleographers.  ...  Leveraging over recent progresses in deep learning, we designed and trained a deep convolutional network achieving an overall accuracy of 96% over the entire dataset, which is one of the highest results  ...  Acknowledgments We thank Debora Benedetto, Elena Bernardi and Riccardo Cecere for their help with the pre-processing steps and the crowd-sourcing application.  ... 
dblp:conf/aiia/FirmaniMNS17 fatcat:zghynsbixzfnthd5so4vrze5e4

PageNet: Page Boundary Extraction in Historical Handwritten Documents [article]

Chris Tensmeyer, Brian Davis, Curtis Wigington, Iain Lee, Bill Barrett
2017 arXiv   pre-print
In PageNet, a Fully Convolutional Network obtains a pixel-wise segmentation which is post-processed into the output quadrilateral region.  ...  We evaluate PageNet on 4 collections of historical handwritten documents and obtain over 94% mean intersection over union on all datasets and approach human performance on 2 of these collections.  ...  Several neural network approaches have been proposed for image segmentation. The Fully Convolution Network (FCN) learns an end-to-end classification function for each pixel in the image [16] .  ... 
arXiv:1709.01618v1 fatcat:sr4tku4uvrg2dd66dkhdw47cjq

A Novel Word Segmentation Method Based on Object Detection and Deep Learning [chapter]

Tomas Wilkinson, Anders Brun
2015 Lecture Notes in Computer Science  
The segmentation of individual words is a crucial step in several data mining methods for historical handwritten documents.  ...  We evaluate its performance using established error metrics, previously used in competitions for word segmentation, and demonstrate its usefulness for a 15th century handwritten document.  ...  Introduction Segmentation of individual words in documents, in particular historical handwritten documents, is a challenging task that is often crucial for further processing and data mining.  ... 
doi:10.1007/978-3-319-27857-5_21 fatcat:tnqeewfpefggpjjr4aftn7vzra

Unsupervised Deep Learning for Handwritten Page Segmentation [article]

Ahmad Droby, Berat Kurar Barakat, Borak Madi, Reem Alaasam, Jihad El-Sana
2021 arXiv   pre-print
Segmenting handwritten document images into regions with homogeneous patterns is an important pre-processing step for many document images analysis tasks.  ...  The network's learned features are used for page segmentation, where patches are classified as main and side text based on the extracted features.  ...  ACKNOWLEDGMENT This research was partially supported by The Frankel Center for Computer Science at Ben-Gurion University.  ... 
arXiv:2101.07487v1 fatcat:qq6i2rezy5dcddvoewrqs5hs2m

An Insight of Script Text Extraction Performance using Machine Learning Techniques

2019 VOLUME-8 ISSUE-10, AUGUST 2019, REGULAR ISSUE  
Firstly, the study presents a survey for various kinds of techniques adopted by the users for extraction of text from image.  ...  Further, it takes a lot of effort and time for script text mining. Therefore, the study investigates various techniques for script text mining viz supervised and unsupervised techniques.  ...  of layers Network (CNN) for page increase performance segmentation of historical document accuracy decreases. 24 Chen.K, (2015) N.Sridevi&P.Subashini (2013) Machine Learning Deep Learning Handwritten  ... 
doi:10.35940/ijitee.a5224.119119 fatcat:rasvnqr5nvevbgjituxs3butsu

WSNet – Convolutional Neural Networkbased Word Spotting for Arabic and English Handwritten Documents

Hanadi Hassen Mohammed, Nandhini Subramanian, Somaya Al-Maadeed, Ahmed Bouridane
2022 TEM Journal  
A Deep learning approach using a novel Convolutional Neural Network is developed for the recognition of the words in historical handwritten documents.  ...  This paper proposes a new convolutional neural network architecture to tackle the problem of word spotting in handwritten documents.  ...  , and deciphering document images, especially handwritten ancient, historical and modern documents.  ... 
doi:10.18421/tem111-33 fatcat:y54qdwikbbfwlakjb2zfzcy64u

Multi-task Layout Analysis for Historical Handwritten Documents Using Fully Convolutional Networks

Yue Xu, Fei Yin, Zhaoxiang Zhang, Cheng-Lin Liu
2018 Proceedings of the Twenty-Seventh International Joint Conference on Artificial Intelligence  
It consists of several sub-processes such as page segmentation, text line segmentation, baseline detection and so on.  ...  The FCN is trained to segment the document image into different regions and detect the center line of each text line by classifying pixels into different categories.  ...  In our work, we proposed a multi-task layout analysis framework based on the fully convolutional network (FCN) for historical handwritten documents.  ... 
doi:10.24963/ijcai.2018/147 dblp:conf/ijcai/XuYZL18 fatcat:tlqazx5ldfatnos5c3ov7u7gca

Table of Contents

2018 2018 16th International Conference on Frontiers in Handwriting Recognition (ICFHR)  
Encoder-Decoder Network for Page Segmentation of Historical Handwritten State University of New York at Buffalo) Documents Into Text Zones and William Barrett (Brigham Young University) Panagiotis Kaddas  ...  and Technology of China), Jun Du (University of Science and Technology of China), and Zi-Rui Wang Watermarking for Security Issue of Handwritten Documents with Fully Convolutional Networks Vinh Loc Cu  ... 
doi:10.1109/icfhr-2018.2018.00004 fatcat:u3d6xtmfsrawhdetd7tkoq2eue

Palmira: A Deep Deformable Network for Instance Segmentation of Dense and Uneven Layouts in Handwritten Manuscripts [article]

Prema Satish Sharan, Sowmya Aitha, Amandeep Kumar, Abhishek Trivedi, Aaron Augustine, Ravi Kiran Sarvadevabhatla
2021 arXiv   pre-print
We also propose a novel deep network Palmira for robust, deformation-aware instance segmentation of regions in handwritten manuscripts.  ...  Handwritten documents are often characterized by dense and uneven layout.  ...  The second contribution is Palmira, a novel deep network architecture for fully automatic region-level instance segmentation of handwritten documents containing dense and uneven layouts.  ... 
arXiv:2108.09436v1 fatcat:ihvf64oggffbbbsy2pn4nikvdy

Full-Page Text Recognition: Learning Where to Start and When to Stop [article]

Bastien Moysset, Christopher Kermorvant, Christian Wolf
2017 arXiv   pre-print
Localization of the text lines is based on regressions with Fully Convolutional Neural Networks and Multidimensional Long Short-Term Memory as contextual layers.  ...  Text line detection and localization is a crucial step for full page document analysis, but still suffers from heterogeneity of real life documents.  ...  In: Workshop on Historical Document and ground-truth alignment of handwritten documents. In: Int.  ... 
arXiv:1704.08628v1 fatcat:wzpw4bxsnvbu7bpulhp765cubi
« Previous Showing results 1 — 15 out of 589 results