Filters








4,512 Hits in 5.6 sec

Deep Structured Output Learning for Unconstrained Text Recognition [article]

Max Jaderberg, Karen Simonyan, Andrea Vedaldi, Andrew Zisserman
2015 arXiv   pre-print
We develop a representation suitable for the unconstrained recognition of words in natural images: the general case of no fixed lexicon and unknown length.  ...  The resulting model is a more accurate system on standard real-world text recognition benchmarks than character prediction alone, setting a benchmark for systems that have not been trained on a particular  ...  We gratefully acknowledge the support of NVIDIA Corporation with the donation of the GPUs used for this research.  ... 
arXiv:1412.5903v5 fatcat:xm54csurlrgrbnxrtv3l3kitie

Visual attention models for scene text recognition [article]

Suman K.Ghosh, Ernest Valveny, Andrew D. Bagdanov
2017 arXiv   pre-print
We validate the performance of our approach on standard SVT and ICDAR'03 scene text datasets, showing state-of-the-art performance in unconstrained text recognition.  ...  In this paper we propose an approach to lexicon-free recognition of text in scene images. Our approach relies on a LSTM-based soft visual attention model learned from convolutional features.  ...  Visual attention for scene text recognition Our recognition approach is based on an encoderdecoder framework for sequence to sequence learning.  ... 
arXiv:1706.01487v1 fatcat:454sz457ingjhb4qwh4otz72em

Sub-sampling Approach for Unconstrained Arabic Scene Text Analysis by Implicit Segmentation based Deep Learning Classifier

Saad Bin Ahmed, Zainab Malik, Muhammad Imran Razzak, Rubiyah Yusof
2019 Global Journal of Computer Science and Technology  
This paper presents a novel contribution and suggests the solution for cursive scene text analysis notably recognition of Arabic scene text appeared in the unconstrained environment.  ...  The deep learning architecture is presented by considering the complexity of the Arabic script. The conducted experiments present 96.81% accuracy at the character level.  ...  Introduction he research on unconstrained scene text recognition is gaining momentum for few years.  ... 
doi:10.34257/gjcstdvol19is1pg7 fatcat:bxhvfwrqbbdpxdpu3ejj2ej3jq

Recursive Recurrent Nets with Attention Modeling for OCR in the Wild

Chen-Yu Lee, Simon Osindero
2016 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
We present recursive recurrent neural networks with attention modeling (R 2 AM) for lexicon-free optical character recognition in natural scene images.  ...  We validate our method with state-of-the-art performance on challenging benchmark datasets: Street View Text, IIIT5k, ICDAR and Synth90k.  ...  Acknowledgements The authors thank Jack Culpepper and Cyprien Noel for valuable discussion, and also thank Flickr Vision and Machine Learning Team for supporting the project.  ... 
doi:10.1109/cvpr.2016.245 dblp:conf/cvpr/LeeO16 fatcat:fx6z5a2bxjf3nclkrvbop2a6de

Recursive Recurrent Nets with Attention Modeling for OCR in the Wild [article]

Chen-Yu Lee, Simon Osindero
2016 arXiv   pre-print
We present recursive recurrent neural networks with attention modeling (R^2AM) for lexicon-free optical character recognition in natural scene images.  ...  We validate our method with state-of-the-art performance on challenging benchmark datasets: Street View Text, IIIT5k, ICDAR and Synth90k.  ...  Acknowledgements The authors thank Jack Culpepper and Cyprien Noel for valuable discussion, and also thank Flickr Vision and Machine Learning Team for supporting the project.  ... 
arXiv:1603.03101v1 fatcat:6fjtkbzvavczbgl7r2seuhzcse

Unconstrained online handwritten Uyghur word recognition based on recurrent neural networks and connectionist temporal classification

Mayire Ibrayim, Wujiahematiti Simayi, Askar Hamdulla
2021 International Journal of Biometrics (IJBM)  
Recognition results are evaluated by calculating the Levenshtein-edit distance and 14.73% character error rate CER on test set of 3,600 samples for 900 word classes has been observed without help of any  ...  An online handwritten word dataset has been established and used for model training and evaluation in writer independent manner.  ...  The second author is grateful to National Laboratory of Pattern Recognition of CASIA (Institute of Automation, Chinese Academy of Sciences) for providing the excellent study and experiment environment.  ... 
doi:10.1504/ijbm.2021.112216 fatcat:js2wjem7hzes3bqugxjf6qinfe

Reading Scene Text in Deep Convolutional Sequences [article]

Pan He, Weilin Huang, Yu Qiao, Chen Change Loy, Xiaoou Tang
2015 arXiv   pre-print
We develop a Deep-Text Recurrent Network (DTRN) that regards scene text reading as a sequence labelling problem.  ...  Codes for the DTRN will be available.  ...  RNN has shown strong capability for learning meaningful structure from an ordered sequence.  ... 
arXiv:1506.04395v2 fatcat:zwa65kok6rbzjpyxiwlhqokx2i

Reading Scene Text in Deep Convolutional Sequences

Pan He, Weilin Huang, Yu Qiao, Chen Loy, Xiaoou Tang
2016 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
We develop a Deep-Text Recurrent Network (DTRN)that regards scene text reading as a sequence labelling problem.  ...  Our model has a number of appealing properties in comparison to existing scene text recognition methods: (i) It can recognise highly ambiguous words by leveraging meaningful context information, allowing  ...  RNN has shown strong capability for learning meaningful structure from an ordered sequence.  ... 
doi:10.1609/aaai.v30i1.10465 fatcat:itw3iqu2o5fkvestpldanqqpqe

An End-to-End Trainable Neural Network for Image-based Sequence Recognition and Its Application to Scene Text Recognition [article]

Baoguang Shi and Xiang Bai and Cong Yao
2015 arXiv   pre-print
Compared with previous systems for scene text recognition, the proposed architecture possesses four distinctive properties: (1) It is end-to-end trainable, in contrast to most of the existing algorithms  ...  recognition tasks. (4) It generates an effective yet much smaller model, which is more practical for real-world application scenarios.  ...  [14] used mid-level features for scene text recognition.  ... 
arXiv:1507.05717v1 fatcat:ix7j6onmpvbr3kcryv3g5orrk4

End to End Recognition System for Recognizing Offline Unconstrained Vietnamese Handwriting [article]

Anh Duc Le, Hung Tuan Nguyen, Masaki Nakagawa
2019 arXiv   pre-print
The model composes of two parts: a DenseNet for extracting invariant features, and a Long Short-Term Memory network (LSTM) with an attention model incorporated for generating output text (LSTM decoder)  ...  This result is competitive with the handwriting recognition system provided by Google in the Vietnamese Online Handwritten Text Recognition competition.  ...  by recent successes in deep learning.  ... 
arXiv:1905.05381v1 fatcat:7e7vqaaqevfzbpnbufhxr44kgy

Practical License Plate Recognition in Unconstrained Surveillance Systems with Adversarial Super-Resolution [article]

Younkwan Lee, Jiwon Jun, Yoojin Hong, Moongu Jeon
2019 arXiv   pre-print
In this paper, we propose a novel license plate recognition method to handle unconstrained real world traffic scenes.  ...  Combined with a deep convolutional network based on VGG-net, our method provides simple but reasonable training procedure.  ...  Adversarial Learning The generative adversarial network (GAN) (Goodfellow et al., 2014; Radford et al., 2015; Radford et al., 2015) is an amazing solution for training deep neural network of generative  ... 
arXiv:1910.04324v1 fatcat:qkl5kakonjhj7lqqlrjbawmha4

An End-to-End Recognition System for Unconstrained Vietnamese Handwriting

Anh Duc Le, Hung Tuan Nguyen, Masaki Nakagawa
2019 SN Computer Science  
The encoder is based on DenseNet for extracting invariant features. The LSTM-based decoder with an attention model incorporated generates output text.  ...  s system, and the system provided by Google in the Vietnamese Online Handwritten Text Recognition competition.  ...  at word level motivated by recent successes in deep learning.  ... 
doi:10.1007/s42979-019-0001-4 fatcat:zniv2vuy3zcirb37b6zput6gvq

Sequence to sequence learning for unconstrained scene text recognition [article]

Ahmed Mamdouh A. Hassanien
2016 arXiv   pre-print
In this work we present a state-of-the-art approach for unconstrained natural scene text recognition.  ...  We show that the LSTM can dramatically reduce such errors and achieve state-of-the-art accuracy in the task of unconstrained natural scene text recognition.  ...  Motaz Abdlewahab for giving me the chance to start my research career. I learned a lot from them both.  ... 
arXiv:1607.06125v1 fatcat:s6heiehojrcujksjfmn7cbcaka

Towards Accurate Handwritten Word Recognition for Hindi and Bangla [chapter]

Kartik Dutta, Praveen Krishnan, Minesh Mathew, C. V. Jawahar
2018 Communications in Computer and Information Science  
Building accurate lexicon free handwritten text recognizers for Indic languages is a challenging task, mostly due to the inherent complexities in Indic scripts in addition to the cursive nature of handwriting  ...  inclusion of spatial transformer layer to learn a model invariant to geometric distortions present in handwriting.  ...  The authors would also like to thank Oishika, Sounak and Sreya for their help in verifying the results for Bangla.  ... 
doi:10.1007/978-981-13-0020-2_41 fatcat:oaqatsafsjchvkljpzm2kas4vm

Research and Analysis of Scene Text Detection and Recognition Technology Based on Deep Learning

Yanju Liu, Xinhai Yi, Yange Li, Huiyu Zhang, Yanzhong Liu
2022 Innovative Computing Information and Control Express Letters, Part B: Applications  
With the development of deep learning technology in the field of computer vision, there are breakthroughs in scene text detection and text recognition technology.  ...  In this paper, the basic concept of the problem is introduced, the scene text detection and text recognition technology are deeply studied from the perspective of deep learning, and the method and regression  ...  Text recognition scene technology is important for scene understanding. It is still the most challenging problem for text recognition in an unconstrained environment.  ... 
doi:10.24507/icicelb.13.04.363 fatcat:ugihkbuu5vf63pvk36al4vzktm
« Previous Showing results 1 — 15 out of 4,512 results