Filters








688 Hits in 13.0 sec

Enhance to Read Better: A Multi-Task Adversarial Network for Handwritten Document Image Enhancement [article]

Sana Khamekhem Jemni and Mohamed Ali Souibgui and Yousri Kessentini and Alicia Fornés
2021 arXiv   pre-print
In this paper, we propose an end to end architecture based on Generative Adversarial Networks (GANs) to recover the degraded documents into a clean and readable form.  ...  the generated document image to be more readable.  ...  Those are training deep learning models, mainly Convolutional Neural Networks (CNNs) and Generative Adversarial Networks (GANs), to learn the parameters for a direct mapping of any degraded document image  ... 
arXiv:2105.12710v2 fatcat:oqri452ocjeqraj4g6t2yxjrpu

A Survey on Deep learning based Document Image Enhancement [article]

Zahra Anvari, Vassilis Athitsos
2022 arXiv   pre-print
Document image enhancement plays a crucial role as a pre-processing step in many automated document analysis and recognition tasks such as character recognition.  ...  With recent advances in deep learning, many methods are proposed to enhance the quality of these document images.  ...  Super-resolution task Very small texts in a document image are often hard to read and very challenging to character recognition methods.  ... 
arXiv:2112.02719v4 fatcat:sznkn6vkr5fabag2pmaff6bhky

Handwriting Recognition in Low-resource Scripts using Adversarial Learning [article]

Ayan Kumar Bhunia, Abhirup Das, Ankan Kumar Bhunia, Perla Sai Raj Kishore, Partha Pratim Roy
2019 arXiv   pre-print
We record results for varying training data sizes, and observe that our enhanced network generalizes much better in the low-data regime; the overall word-error rates and mAP scores are observed to improve  ...  We propose the Adversarial Feature Deformation Module (AFDM) that learns ways to elastically warp extracted features in a scalable manner.  ...  The technology is applicable in postal automation, bank cheque processing, digitization of handwritten documents, as well as reading aid for the visually handicapped.  ... 
arXiv:1811.01396v5 fatcat:xp3emb4whrh7jasluh3wv3ffce

Text-DIAE: A Self-Supervised Degradation Invariant Autoencoders for Text Recognition and Document Enhancement [article]

Mohamed Ali Souibgui, Sanket Biswas, Andres Mafla, Ali Furkan Biten, Alicia Fornés, Yousri Kessentini, Josep Lladós, Lluis Gomez, Dimosthenis Karatzas
2022 arXiv   pre-print
In this paper, we propose a Text-Degradation Invariant Auto Encoder (Text-DIAE), a self-supervised model designed to tackle two tasks, text recognition (handwritten or scene-text) and document image enhancement  ...  Finally, we demonstrate that our method surpasses the state-of-the-art in existing supervised and self-supervised settings in handwritten and scene text recognition and document image enhancement.  ...  Fig. 3 . 3 For each task, given an unlabeled image I (eg. a cropped handwritten text, cropped scene text or a scanned document image), we use a function φ to map I to a degraded form.  ... 
arXiv:2203.04814v4 fatcat:xjcutahg6vbj7plc5jbbkkgyqm

Deep Learning for Historical Document Analysis and Recognition—A Survey

Francesco Lombardi, Simone Marinai
2020 Journal of Imaging  
This analysis shows that the latest research is a leap forward since it is not the simple use of recently proposed algorithms to previous problems, but novel tasks and novel applications of state of the  ...  Rather than just providing a conclusive picture of the current research in the topic we lastly suggest some potential future trends that can represent a stimulus for innovative research directions.  ...  Two main related operations are performed for preprocessing: document image enhancement (including image denoising and image restoration) and document binarization.  ... 
doi:10.3390/jimaging6100110 pmid:34460551 pmcid:PMC8321201 fatcat:nevh2ctshzfwtey4girgjtaftq

Unsupervised Writer Adaptation for Synthetic-to-Real Handwritten Word Recognition [article]

Lei Kang, Marçal Rusiñol, Alicia Fornés, Pau Riba, Mauricio Villegas
2019 arXiv   pre-print
In this paper, we propose an unsupervised writer adaptation approach that is able to automatically adjust a generic handwritten word recognizer, fully trained with synthetic fonts, towards a new incoming  ...  Across these challenging collections, we show that our system is able to maintain its performance, thus, it provides a practical and generic approach to deal with new document collections without requiring  ...  Acknowledgements This work has been partially supported by the European Commission H2020 SME Instrument program, project OMNIUS: SaaS platform for automated categorization and mapping of digitized documents  ... 
arXiv:1909.08473v1 fatcat:asn53clgebei7id2vvdzxpkn4q

Unsupervised Adaptation for Synthetic-to-Real Handwritten Word Recognition

Lei Kang, Marcal Rusinol, Alicia Fornes, Pau Riba, Mauricio Villegas
2020 2020 IEEE Winter Conference on Applications of Computer Vision (WACV)  
In this paper, we propose an unsupervised writer adaptation approach that is able to automatically adjust a generic handwritten word recognizer, fully trained with synthetic fonts, towards a new incoming  ...  Across these challenging collections, we show that our system is able to maintain its performance, thus, it provides a practical and generic approach to deal with new document collections without requiring  ...  Acknowledgements This work has been partially supported by the European Commission H2020 SME Instrument program, project OMNIUS: SaaS platform for automated categorization and mapping of digitized documents  ... 
doi:10.1109/wacv45572.2020.9093392 dblp:conf/wacv/KangRFRV20 fatcat:oifmc2ljora25ob35xfiwgfaem

Handwritten text generation and strikethrough characters augmentation [article]

Alex Shonenkov, Denis Karachev, Max Novopoltsev, Mark Potanin, Denis Dimitrov, Andrey Chertok
2021 arXiv   pre-print
We apply a novel augmentation that simulates strikethrough text (HandWritten Blots) and a handwritten text generation method based on printed text (StackMix), which proved to be very effective in HTR tasks  ...  Because these data augmentation techniques are independent of the network used, they could also be applied to enhance the performance of other networks and approaches to HTR.  ...  Authors use Generative Adversarial Networks (GAN) (called ScrabbleGAN) to generate training data. 12 Scrab-bleGAN follows a semi-supervised approach to synthesize handwritten text images that are versatile  ... 
arXiv:2112.07395v1 fatcat:5kwex6tsm5az5iz5jejhr4vvwq

When Counting Meets HMER: Counting-Aware Network for Handwritten Mathematical Expression Recognition [article]

Bohan Li, Ye Yuan, Dingkang Liang, Xiao Liu, Zhilong Ji, Jinfeng Bai, Wenyu Liu, Xiang Bai
2022 arXiv   pre-print
To alleviate this problem, we propose an unconventional network for HMER named Counting-Aware Network (CAN), which jointly optimizes two tasks: HMER and symbol counting.  ...  Recently, most handwritten mathematical expression recognition (HMER) methods adopt the encoder-decoder networks, which directly predict the markup sequences from formula images with the attention mechanism  ...  HME100K Dataset [37] is a real scene handwritten mathematical expression dataset, consisting of 74,502 images for training and 24,607 images for testing.  ... 
arXiv:2207.11463v1 fatcat:sb3wu25dtrejbbmwwvp6e3oeke

Learning Neural Textual Representations for Citation Recommendation

Binh Thanh Kieu, Inigo Jauregi Unanue, Son Bao Pham, Hieu Xuan Phan, Massimo Piccardi
2021 2020 25th International Conference on Pattern Recognition (ICPR)  
Network and Task Embedding for Multi-Task Regression Problems.pdf Learning from Learners: Adapting Reinforcement Learning Agents to be Competitive in a Card Game.  ...  CH3.2 A Hierarchical Multi-Task Approach to Gastrointestinal Image Analysis DAY 2 -Jan 13, 2021 Live Zhipeng Luo et al.  ... 
doi:10.1109/icpr48806.2021.9412725 fatcat:3vge2tpd2zf7jcv5btcixnaikm

Derin Öğrenme Araştırma Alanlarının Literatür Taraması

M. Mutlu Yapıcı, Adem Tekerek, Nurettin Topaloğlu
2019 Gazi Mühendislik Bilimleri Dergisi  
In this study, it is presented important knowledge to guide about DL models and challenging topics that can be used in DL for researchers.  ...  In the present day, Deep learning methods have reached better results than humans in object recognition.  ...  A new training method to enhance Deep Convolutional Neural Networks (DCNN) in handwritten Chinese character recognition was proposed by Yang et al. [125] .  ... 
doi:10.30855/gmbd.2019.03.01 fatcat:2sv7dg7elrfqppcjx5otzmb7pi

An enhanced binarization framework for degraded historical document images

Wei Xiong, Lei Zhou, Ling Yue, Lirong Li, Song Wang
2021 EURASIP Journal on Image and Video Processing  
In this paper, we present our winning algorithm in ICFHR 2018 competition on handwritten document image binarization (H-DIBCO 2018), which is based on background estimation and energy minimization.  ...  Second, we perform Laplacian energy-based segmentation on the compensated document images. Finally, we implement post-processing to preserve text stroke connectivity and eliminate isolated noise.  ...  [47] formulate binarization as an image-to-image generation task, using conditional generative adversarial networks (cGANs) to solve multi-scale information combination problems in binarization tasks  ... 
doi:10.1186/s13640-021-00556-4 fatcat:aawmkhjf3ngs3bb3zbyyi4xz5q

Detection Method for Classifying Malicious Firmware

David Noever, Samantha E. Miller Noever
2021 International journal of network security and its applications  
This work converts the binary headers of 40,000 firmware examples from bytes into 1024-pixel thumbnail images to train a deep neural network.  ...  A malicious firmware update may prove devastating to the embedded devices both that make up the Internet of Things (IoT) and that typically lack the same security verifications now applied to full operating  ...  "An analysis of generative adversarial networks and variants for image synthesis on MNIST dataset."  ... 
doi:10.5121/ijnsa.2021.13601 fatcat:m2uopqqovngdzee4zbndpztocq

Attention-Guided Answer Distillation for Machine Reading Comprehension [article]

Minghao Hu, Yuxing Peng, Furu Wei, Zhen Huang, Dongsheng Li, Nan Yang, Ming Zhou
2018 arXiv   pre-print
We first demonstrate that vanilla knowledge distillation applied to answer span prediction is effective for reading comprehension systems.  ...  Besides, existing approaches are also vulnerable to adversarial attacks.  ...  Acknowledgments We would like to thank Pranav Rajpurkar for his help with SQuAD submissions. This work is funded by National Key R&D Program of China (No. 0708063216003).  ... 
arXiv:1808.07644v4 fatcat:wcnhvuat5veelomvmj5rntwmre

A Survey of Unsupervised Deep Domain Adaptation [article]

Garrett Wilson, Diane J. Cook
2020 arXiv   pre-print
Deep learning has produced state-of-the-art results for a variety of tasks.  ...  As a complement to this challenge, single-source unsupervised domain adaptation can handle situations where a network is trained on labeled data from a source domain and unlabeled data from a related but  ...  Multi-adversarial domain adaptation (MADA) [183] combines adversarial domain-invariant feature learning with ensemble methods for the purpose of better handling multi-modal data.  ... 
arXiv:1812.02849v3 fatcat:paefg5cywbe3tjsp6dffnwkvxy
« Previous Showing results 1 — 15 out of 688 results