52 Hits in 4.2 sec

Deep Network with Pixel-Level Rectification and Robust Training for Handwriting Recognition

Shanyu Xiao, Liangrui Peng, Ruijie Yan, Shengjin Wang
2020 SN Computer Science  
In order to learn invariant feature representations for handwriting, we propose a novel method to incorporate pixel-level rectification into a CNN-and RNNbased recognizer.  ...  Offline handwriting recognition is a well-known challenging task in the optical character recognition field due to the difficulty caused by various unconstrained handwriting styles and limited training  ...  In this paper, we propose a deep network with pixellevel rectification and robust training methods to address the above two problems for handwriting recognition.  ... 
doi:10.1007/s42979-020-00133-y fatcat:po46k6cnqncvrlyi7qefvvn7p4

Robust End-to-End Offline Chinese Handwriting Text Page Spotter with Text Kernel [article]

Zhihao Wang, Yanwei Yu, Yibo Wang, Haixu Long, Fazheng Wang
2021 arXiv   pre-print
Without any language model, the correct rates are 99.12% and 94.27% for line-level recognition, and 99.03% and 94.20% for page-level recognition, respectively.  ...  and improves the robustness of the system.  ...  [6] proposed a deep network with Pixel-Level Rectification to integrate pixel-level rectification into CNN and RNN-based recognizers.  ... 
arXiv:2107.01547v1 fatcat:xq4d7xiirrg4zofkirf3cm7sqa

TextAdaIN: Paying Attention to Shortcut Learning in Text Recognizers [article]

Oren Nuriel, Sharon Fogel, Ron Litman
2021 arXiv   pre-print
Leveraging the characteristics of convolutional layers, neural networks are extremely effective for pattern recognition tasks.  ...  TextAdaIN achieves state-of-the-art results on standard handwritten text recognition benchmarks. Additionally, it generalizes to multiple architectures and to the domain of scene text recognition.  ...  Models trained on handwriting and scene text datasets are trained for 200k iterations and 600k iterations, respectively.  ... 
arXiv:2105.03906v2 fatcat:v4pel3x4d5cttiivvrjdccgsgi

Recognition of Off-line Kannada Handwritten Characters by Deep Learning using Capsule Network

2019 International Journal of Engineering and Advanced Technology  
Machine learning and Deep learning approaches to the problem have yielded acceptable results throughout, yet there is still room for improvement. off-line Kannada handwritten character recognition is another  ...  We also carefully consider the drawbacks of CNN and its impact on the problem statement, which are solved with the usage of Capsule Networks.  ...  The last layer of the network is convolutional and the spatial information is place-coded for low-level capsules and rate-coded for higher level capsules in the hierarchy.  ... 
doi:10.35940/ijeat.f8726.088619 fatcat:zjiloymhvvh33oqkpazq7xqsfy

Gender and Handedness Prediction from Offline Handwriting Using Convolutional Neural Networks

Ángel Morera, Ángel Sánchez, José Francisco Vélez, Ana Belén Moreno
2018 Complexity  
Our research was carried out on two public handwriting databases: the IAM dataset containing English texts and the KHATT one with Arabic texts.  ...  debugging of these deep architectures when handling related handwriting problems.  ...  Acknowledgments This research has been supported by the Spanish Ministerio de Economía y Competitividad (MINECO), under the Projects TIN2014-57458-R and TIN2017-85221-R. Complexity 13  ... 
doi:10.1155/2018/3891624 fatcat:rpixbvem4zh7hevpyru6gpyhsm

Text Detection and Recognition in the Wild: A Review [article]

Zobeir Raisi, Mohamed A. Naiel, Paul Fieguth, Steven Wardell, John Zelek
2020 arXiv   pre-print
The current state-of-the-art scene text detection and/or recognition methods have exploited the witnessed advancement in deep learning architectures and reported a superior accuracy on benchmark datasets  ...  , but also presenting the results of conducting extensive experiments using a unified evaluation framework that assesses pre-trained models of the selected methods on challenging cases, and applies the  ...  Acknowledgements The authors would like to thank the Ontario Centres of Excellence (OCE) -Voucher for Innovation and Productivity II (VIP II) -Canada program, and ATS Automation Tooling Systems Inc., Cambridge  ... 
arXiv:2006.04305v2 fatcat:paccfprli5arbj4ggfx5z3hrve

Continuous Offline Handwriting Recognition using Deep Learning Models [article]

Jorge Sueiras
2021 arXiv   pre-print
For this purpose, we have proposed a new recognition model based on integrating two types of deep learning architectures: convolutional neural networks (CNN) and sequence-to-sequence (seq2seq) models,  ...  Additionally, extensive experimentation of the proposed model for the continuous problem has been carried out to determine its robustness to changes in parameterization.  ...  The case of handwriting recognition is not an exception and several deep architectures has been proposed for the word and line recognition problems with significant improvements in the results, as shown  ... 
arXiv:2112.13328v1 fatcat:xkcdw7c2rngd7jsaixsfosqzc4

Scene Text Detection and Recognition: The Deep Learning Era [article]

Shangbang Long, Xin He, Cong Yao
2020 arXiv   pre-print
With the rise and development of deep learning, computer vision has been tremendously transformed and reshaped.  ...  This survey is aimed at summarizing and analyzing the major changes and significant progresses of scene text detection and recognition in the deep learning era.  ...  One conjecture is that the robustness of models is related to the internal operating mechanism of deep neural networks.  ... 
arXiv:1811.04256v5 fatcat:vhtpriukobcu7cwikh6me2wuwm

Text is Text, No Matter What: Unifying Text Recognition using Knowledge Distillation [article]

Ayan Kumar Bhunia, Aneeshan Sain, Pinaki Nath Chowdhury, Yi-Zhe Song
2021 arXiv   pre-print
The challenging nature of the very problem however dictated a fragmentation of research efforts: Scene Text Recognition (STR) that deals with text in everyday scenes, and Handwriting Text Recognition (  ...  In this paper, for the first time, we argue for their unification -- we aim for a single model that can compete favourably with two separate state-of-the-art STR and HTR models.  ...  Related Works Text Recognition: With the inception of deep learning, Jaderberg et al. [27, 26] introduced a dictionary-based text recognition framework employing deep networks.  ... 
arXiv:2107.12087v2 fatcat:z2ezlupwyrh5xjykeh6bdkjwg4

2D Positional Embedding-based Transformer for Scene Text Recognition

Zobeir Raisi, Mohamed A. Naiel, Paul Fieguth, Steven Wardell, John Zelek
2021 Journal of Computational Vision and Imaging Systems  
Recent state-of-the-art scene text recognition methods are primarily based on Recurrent Neural Networks (RNNs), however, these methods require one-dimensional (1D) features and are not designed for recognizing  ...  In this paper, we leverage a Transformer-based architecture for recognizing both regular and irregular text-in-the-wild images.  ...  Acknowledgments The authors would like to thank the Ontario Centres of Excellence (OCE) -Voucher for Innovation and Productivity II (VIP II) -Canada program, and ATS Automation Tooling Systems Inc., Cambridge  ... 
doi:10.15353/jcvis.v6i1.3533 fatcat:q5hid4xetrf7fnmwzeocjapmq4


Musab COŞKUN, Özal YILDIRIM, Ayşegül UÇAR, Yakup DEMIR
2017 European Journal of Technic  
Deep learning has shown great successes in many domains such as handwriting recognition, image recognition, object detection etc.  ...  We revisited the concepts and mechanisms of typical deep learning algorithms such as Convolutional Neural Networks, Recurrent Neural Networks, Restricted Boltzmann Machine, and Autoencoders.  ...  For example, these networks can transform an image containing 28x28 grid of pixels into a representation with only 30 numbers.  ... 
doi:10.23884/ejt.2017.7.2.11 fatcat:l3oygb3ljbgzvca7kodmziu3me

First Steps Toward Camera Model Identification With Convolutional Neural Networks

Luca Bondi, Luca Baroffio, David Guera, Paolo Bestagini, Edward J. Delp, Stefano Tubaro
2017 IEEE Signal Processing Letters  
generalize to camera models never used for training.  ...  Specifically, we propose a data-driven algorithm based on convolutional neural networks, which learns features characterizing each camera model directly from the acquired pictures.  ...  and handwriting recognition [19] .  ... 
doi:10.1109/lsp.2016.2641006 fatcat:nywbfa43ozcnrcuug4cxjkhu2q

DocUNet: Document Image Unwarping via a Stacked U-Net

Ke Ma, Zhixin Shu, Xue Bai, Jue Wang, Dimitris Samaras
2018 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition  
The network is trained on this dataset with various data augmentations to improve its generalization ability. We further create a comprehensive benchmark 1 that covers various real-world conditions.  ...  Capturing document images is a common way for digitizing and recording physical documents due to the ubiquitousness of mobile cameras.  ...  This work was supported by a gift from Adobe, Partner University Fund, and the SUNY2020 Infrastructure Transportation Security Center.  ... 
doi:10.1109/cvpr.2018.00494 dblp:conf/cvpr/MaSBWS18 fatcat:zpcvr3rjkrgg3kk5omq35c23si

Mobile Vision-Based Sketch Recognition with SPARK [article]

Jeffrey Browne, Timothy Sherwood
2012 Sketch-Based Interfaces and Modeling  
for building mobile, image-based sketch recognition applications.  ...  In this paper we explore bringing the benefits of sketch capture and recognition to traditional surfaces through a common smart-phone with the Sketch Practically Anywhere Recognition Kit (SPARK), a framework  ...  Acknowledgments The authors would like to thank Steven Boyd for implementation help, and Dr. Plimmer's sketch research group at the University of Auckland for providing the labeled stroke data.  ... 
doi:10.2312/sbm/sbm12/087-096 fatcat:m5gmqbkjffdm5fow7k76s2km54

Text Recognition in the Wild: A Survey [article]

Xiaoxue Chen, Lianwen Jin, Yuanzhi Zhu, Canjie Luo, Tianwei Wang
2020 arXiv   pre-print
In recent years, with the rise and development of deep learning, numerous methods have shown promising in terms of innovation, practicality, and efficiency.  ...  This paper aims to (1) summarize the fundamental problems and the state-of-the-art associated with scene text recognition; (2) introduce new insights and ideas; (3) provide a comprehensive review of publicly  ...  Compared with realistic datasets, multi-level annotation information (i.e., word-level, character-level and pixel-level) can be easily obtained during synthesizing, which can be used to train data-hungry  ... 
arXiv:2005.03492v3 fatcat:rmzmavxylnf6rbp52lje2mrgiy
« Previous Showing results 1 — 15 out of 52 results