17 Hits in 9.6 sec

Identifying Table Structure in Documents using Conditional Generative Adversarial Networks [article]

Nataliya Le Vine, Claus Horn, Matthew Zeigenfuse, Mark Rowan
2020 arXiv   pre-print
The approach is easily adaptable to different table configurations and requires small data set sizes for training.  ...  In many industries, as well as in academic research, information is primarily transmitted in the form of unstructured documents (this article, for example).  ...  is thus directly related to the accuracy of the skew angle estimation.  ... 
arXiv:2001.05853v1 fatcat:vrweadv4p5d47p67ckx3stn5wu

TMIXT: A process flow for Transcribing MIXed handwritten and machine-printed Text

Fady Medhat, Mahnaz Mohammadi, Sardar Jaf, Chris G. Willcocks, Toby P. Breckon, Peter Matthews, Andrew Stephen McGough, Georgios Theodoropoulos, Boguslaw Obara
2018 2018 IEEE International Conference on Big Data (Big Data)  
of scanned documents which need to be processed in a finite time.  ...  However, this problem is exacerbated both by the volume, in terms of scanned documents and the complexity of the pages, which need to be processed.  ...  For example, in Profile Projection (PP) analysis, the image is projected to a single vector and further analysis is applied to estimate the skewness angle.  ... 
doi:10.1109/bigdata.2018.8622136 dblp:conf/bigdataconf/MedhatMJWBMMTO18 fatcat:wtvr6sclrrcgvo3b2o6x74m74i

Print-Scan Resilient Text Image Watermarking Based on Stroke Direction Modulation for Chinese Document Authentication

L. Tan, X. Sun, G. Sun
2012 Radioengineering  
Experimental results show that our technique attains high detection accuracy against distortions resulting from print-scan operations, good quality photocopies and benign attacks in accord with the future  ...  During the embedding phase, the angle of rotatable strokes are quantized to embed the bits.  ...  Fig. 9 . 9 Skew estimation: (a) scanned image, (b) centroid labeling, (c) HT of the centroid image, (d) Hough-based line finding. Fig. 11 . 2 112 Various types of attacks.  ... 
doaj:685cbb187f164ad09ca9cff908902ac3 fatcat:2sdcau3fzjednkuawk5flhcame

Automatic Transcription of Organ Tablature Music Notation with Deep Neural Networks

Daniel Schneider, Nikolaus Korfhage, Markus Mühling, Peter Lüttig, Bernd Freisleben
2021 Transactions of the International Society for Music Information Retrieval  
Overall, our approach achieves an accuracy of 97.2% and 99.3% correctly recognized bars, depending on whether note pitch and rest characters or note duration and special characters are considered, respectively  ...  In this paper, we present a deep learning approach to automatically recognize organ tablature notation in scanned documents and transcribe it to modern music notation.  ...  This includes image deskewing and segmentation into rows and individual staves. Deskewing The quality of scans of old documents can vary significantly.  ... 
doi:10.5334/tismir.77 fatcat:jgd7mxj255fufpbxezfg22xi3q

Optical Recognition of Handwritten Logic Formulas Using Neural Networks

Vaios Ampelakiotis, Isidoros Perikos, Ioannis Hatzilygeroudis, George Tsihrintzis
2021 Electronics  
The final accuracy achieved is 90.13%. The general methodology followed consists of two stages: the image processing and the NN design and training.  ...  algorithm, optimized by the Adam update rule, which was proved to be the best, using a trainset of 16,750 handwritten image samples of 28 × 28 each and a testset of 7947 samples.  ...  , calculate the new angle, and warp affine the image to deskew it.  ... 
doi:10.3390/electronics10222761 fatcat:uvgox26qv5e7jo72ciwomv4lw4

On the Farey sequence and its augmentation for applications to image analysis

Sanjoy Pratihar, Partha Bhowmick
2017 International Journal of Applied Mathematics and Computer Science  
To assert its merit, we show its use in two applications—one in polygonal approximation of digital curves and the other in skew correction of engineering drawings in document images.  ...  straight lines—often required to solve many image-analytic problems—can be made fast and efficient through an appropriate AFT-based tool.  ...  ., Pratihar, S. and Bhowmick, P. (2010  ... 
doi:10.1515/amcs-2017-0045 fatcat:2cu7jwgd4rgrbdpucisn7xva4m

OCR4all – An Open-Source Tool Providing a (Semi-)Automatic OCR Workflow for Historical Printings [article]

Christian Reul, Dennis Christ, Alexander Hartelt, Nico Balbach, Maximilian Wehner, Uwe Springmann, Christoph Wick, Christine Grundig, Andreas Büttner, Frank Puppe
2019 arXiv   pre-print
Further on, extensive configuration capabilities are provided to set the degree of automation of the workflow and to make adaptations to the carefully selected default parameters for specific printings  ...  Optical Character Recognition (OCR) on historical printings is a challenging task mainly due to the complexity of the layout and the highly variant typography.  ...  Finally, our gratitude is due to everyone who supported the planning, implementation, evaluation, distribution, and presentation of OCR4all including Sophia Beckenbauer, Kevin Chadbourne, Björn Eyselein  ... 
arXiv:1909.04032v1 fatcat:czzg6o6i5baxdcnsc2cacm5xmy

Illustrated Book Study: Digital Conversion Requirements of Printed Illustrations [chapter]

Anne R. Kenney, Louis H. Sharpe, Barbara Berger
1998 Lecture Notes in Computer Science  
Deskewing. Removal of small angle rotation of the image content relative to the scanning axes. • Inverse Halftoning.  ...  This view is based on the psycho-visual experience of the reader rather than any feature associated with the source document.  ...  Scope This document describes PEI's halftone illustration detection and de-screening algorithm and gives instructions for using the simulation of the algorithm provided in the software package pei1.250  ... 
doi:10.1007/3-540-49653-x_17 fatcat:n5byitzncnd33ondnnokbvkzja

A neural network approach to online Devanagari handwritten character recognition

Shruthi Kubatur, Maher Sid-Ahmed, Majid Ahmadi
2012 2012 International Conference on High Performance Computing & Simulation (HPCS)  
Many algorithms have been proposed for adapting weights in neural networks that are based on backpropagation algorithm.  ...  Offline Vs Online Handwriting Recognition As discussed, handwriting recognition can be either offline or online depending on whether the input is scanned and digitized copy of handwritten documents or  ...  APPENDIX Matlab code common to training and testing phase  ... 
doi:10.1109/hpcsim.2012.6266913 dblp:conf/ieeehpcs/KubaturSA12 fatcat:dutwfsotmrdc3egagm6wxx4oy4

Fast registration of tabular document images using the Fourier-Mellin transform

L.A.D. Hutchison, W.A. Barrett
First International Workshop on Document Image Analysis for Libraries, 2004. Proceedings.  
If the value of the horizontal scale is unrealistic based on the value of the vertical scale, the vertical scale could be used for both.  ...  ) ± 45 • (modulo 180 • ). θ 1 and θ 2 are both at right angles to one of the document axes.  ...  A.1.1 C source code for 1D quadratic parameter fitting The following code illustrates "quadratic parameter fitting" (or "parabolic neighborhood fitting") for improvement of parameter estimates, as described  ... 
doi:10.1109/dial.2004.1263254 dblp:conf/dial/HutchisonB04 fatcat:b5lmz3vcjvcr7mxla5sglygioy

Warping-Based Approach to Offline Handwriting Recognition Warping-Based Approach to Offline Handwriting Recognition

Douglas Kennard, Douglas Kennard, William Barrett, Bryan Morse, Eric Ringger, Dan Olsen, Daniel Zappala, Douglas Kennard
2013 unpublished
Much of the recent HR work has focused on incremental improvements to methods based on Hidden Markov Models (HMMs) and other similar probabilistic approaches.  ...  Even after being scanned into images, only a minute fraction of the existing records can be manually transcribed / indexed with reasonable amounts of time and cost.  ...  The angle estimation uses ink runlengths accumulated into a histogram based on angle bins. Baseline estimation is used to determine whether to pad the top or bottom of the image.  ... 

Automated Semantic Annotation of Historical Catalogues

David Körner, Robert Sablatnig, Markus Diem
For this purpose, an approach based on Maximally Stable Extremal Regions and a subsequent text region grouping is used.  ...  This classification is done using a texture analysis of the word regions based on Gabor filtering.  ...  Acknowledgements I would like to thank my advisors Florian Kleber and Markus Diem for their advice, feedback and most of all their patience.  ... 
doi:10.34726/hss.2020.47722 fatcat:eglt5o7wzbfhjhf6jmz3ggxzmi

A system for optical music recognition and audio synthesis

Matthias Wallner, Horst Eidenberger
There is a particular emphasis on projection-based methods, which have proven highly successful as early as in the 1980s. Of all the works that were carried out, the research of I.  ...  The focus is on the fields of image processing and pattern recognition as well as audio synthesis.  ...  For the actual Deskewing It is an unfavourable property of photographs or even some image scans, that their main objects are often skewed at an arbitrary angle, which provides a suboptimal starting point  ... 
doi:10.34726/hss.2015.25684 fatcat:wwjdjqrwlbeazbxauhuqbw3lyq

Visual Representation Learning for Document Image Recognition [article]

(:Unkn) Unknown, National Technological University Of Athens, National Technological University Of Athens
and subsequently achieve high accuracy.  ...  AlexNet consists of 8 layers (5 convolutional and 3 fully-connected) with high variability in size, which is ideal to showcase the value of adaptive pruning approaches.  ... 
doi:10.26240/heal.ntua.17645 fatcat:acp2tnnfvvc2pntft7kxgzrjbq

Modern Time: Photography and Temporality

Kris Belden-Adams
2009 International Journal of Technology, Knowledge and Society  
Rather than provide a comprehensive, and necessarily incomplete, study of every possible way in which photography can relate to time, this study instead focuses on a number of in-depth  ...  to all commentators on photographywhat exactly is photography's relationship to time, and by extension, to reality?  ...  Skew is corrected mathematically, according to Avery: -Each scene is deskewed by an algorithm that shifts scan lines by a calculated number of pixelsa number dependent on the estimated latitude for the  ... 
doi:10.18848/1832-3669/cgp/v05i03/55994 fatcat:xduqv4o7ovbp5i6hjk2telbsly
« Previous Showing results 1 — 15 out of 17 results