15,950 Hits in 8.1 sec

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

Shangbang Long, Xin He, Cong Yao
2020 arXiv   pre-print
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.  ...  As an important research area in computer vision, scene text detection and recognition has been inescapably influenced by this wave of revolution, consequentially entering the era of deep learning.  ...  Methods before the Deep Learning Era In this section, we take a glance retrospectively at algorithms before the deep learning era.  ... 
arXiv:1811.04256v5 fatcat:vhtpriukobcu7cwikh6me2wuwm

Arabic Scene Text Recognition in the Deep Learning Era: Analysis on A Novel Dataset

Heba Hassan, Ahmed El-Mahdy, Mohamed E. Hussein.
2021 IEEE Access  
Marwan Torki (Alexandria University) for his valuable ideas and suggestions during the planning of this work. His willingness to give his time so generously is very much appreciated.  ...  of Arabic text recognition using a selected set of deep learning-based scene text recognition models to determine which works better with Arabic text.  ...  SCENE TEXT RECOGNITION Before deep learning, scene text recognition was usually performed using hand-crafted features that utilize text characteristics to extract useful features from the image.  ... 
doi:10.1109/access.2021.3100717 fatcat:vktg6bdybjhtnb2f67yovifo4q

Saliency Prediction in the Deep Learning Era: Successes, Limitations, and Future Challenges [article]

Ali Borji
2019 arXiv   pre-print
In this work, I explore the landscape of the field emphasizing on new deep saliency models, benchmarks, and datasets.  ...  Visual saliency models have enjoyed a big leap in performance in recent years, thanks to advances in deep learning and large scale annotated data.  ...  SALIENCY IN THE DEEP LEARNING ERA Benchmarks Benchmarks have been instrumental for advances in computer vision.  ... 
arXiv:1810.03716v3 fatcat:zw3piox3v5evvaucfupzmvjryq

Synthetic-to-Real Unsupervised Domain Adaptation for Scene Text Detection in the Wild [article]

Weijia Wu and Ning Lu and Enze Xie
2020 arXiv   pre-print
Deep learning-based scene text detection can achieve preferable performance, powered with sufficient labeled training data. However, manual labeling is time consuming and laborious.  ...  In this paper, a text self-training (TST) method and adversarial text instance alignment (ATA) for domain adaptive scene text detection are introduced.  ...  Related Work Scene Text Detection Before the era of deep learning, SWT [6] and MSER [25] were two representative algorithms for conventional text detection methods.  ... 
arXiv:2009.01766v1 fatcat:2g44eg4zijdgljd2huhsx4pt3y

Recent Advancements in Text Detection Methods from Natural Scene Images

Shiravale S. S., Sannakki S. S., Rajpurohit V. S.
2020 International journal of engineering research and technology  
Text detection and recognition are two main steps of such text-based applications. Text detection from natural scene images is tedious compared to text detection from document images.  ...  The main purpose of this paper is to highlight available text detection methods with pros and cons, challenges in the text detection process, evaluation parameters as well as recent achievements.  ...  V.III Deep Neural Network Recently deep neural networks are fetching great success in object detection and recognition problems.  ... 
doi:10.37624/ijert/13.6.2020.1344-1352 fatcat:7sfxirdwa5a7hdif3s4p2blzgu

Polygon-free: Unconstrained Scene Text Detection with Box Annotations [article]

Weijia Wu, Enze Xie, Ruimao Zhang, Wenhai Wang, Hong Zhou, Ping Luo
2022 arXiv   pre-print
Although a polygon is a more accurate representation than an upright bounding box for text detection, the annotations of polygons are extremely expensive and challenging.  ...  We hope that PF can provide a new perspective for text detection to reduce the labeling costs.  ...  Related Work Supervised Text Detection Scene text detection has achieved remarkable progress in the deep learning era.  ... 
arXiv:2011.13307v3 fatcat:xwzeylglofckrmarmiy3ru44zu

Content-based Image Retrieval and the Semantic Gap in the Deep Learning Era [article]

Björn Barz, Joachim Denzler
2020 arXiv   pre-print
We conclude that the key problem for the further advancement of semantic image retrieval lies in the lack of a standardized task definition and an appropriate benchmark dataset.  ...  We then apply them to a semantic image retrieval task and find that they perform inferior to much less sophisticated and more generic methods in a setting that requires image understanding.  ...  Impact on the Semantic Gap The previous section outlined the impressive advances of instance retrieval in the deep learning era.  ... 
arXiv:2011.06490v1 fatcat:fgrcgt2jxbdchfe7ts7t6ephcy

Accurate Scene Text Detection via Scale-Aware Data Augmentation and Shape Similarity Constraint

Pengwen Dai, Yang Li, Hua Zhang, Jingzhi Li, Xiaochun Cao
2021 IEEE transactions on multimedia  
Scene text detection has attracted increasing concerns with the rapid development of deep neural networks in recent years.  ...  This paper presents an arbitrary-shape scene text detection method that can achieve better generalization ability and more accurate localization.  ...  RELATED WORK In the era of deep learning, most scene text detection methods can be found in the recent survey [38] .  ... 
doi:10.1109/tmm.2021.3073575 fatcat:r4rgthwbk5cw3boixvko3zbpoi

Deep Learning in the Era of Edge Computing: Challenges and Opportunities [article]

Mi Zhang, Faen Zhang, Nicholas D. Lane, Yuanchao Shu, Xiao Zeng, Biyi Fang, Shen Yan, Hui Xu
2020 arXiv   pre-print
The era of edge computing has arrived.  ...  However, deep learning-based approaches require a large volume of high-quality data to train and are very expensive in terms of computation, memory, and power consumption.  ...  of important computing tasks such as speech recognition [9] , machine translation [1] , object recognition [13] , face detection [23] , sign language translation [5] , and scene understanding [27  ... 
arXiv:2010.08861v1 fatcat:2lrjtm6nj5b7nbnilkc74jhbea

Attentive models in vision: Computing saliency maps in the deep learning era

Marcella Cornia, Davide Abati, Lorenzo Baraldi, Andrea Palazzi, Simone Calderara, Rita Cucchiara, Stefano Ferilli, Francesca Alessandra Lisi
2019 Intelligenza Artificiale  
In the last few years, early models have been replaced by deep learning architectures, that outperform any early approach compared against public datasets.  ...  saliency models, although improved with the semantics learned by human groundthuth.  ...  cues as well, like faces, text, and the horizon.  ... 
doi:10.3233/ia-170033 fatcat:votkkk4skrajhhkzdkuvjp77zu

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  ...  Thus, unlike previous surveys in this field, the objectives of this survey are as follows: first, offering the reader not only a review on the recent advancement in scene text detection and recognition  ...  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

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  ...  In this section, we introduce new insights and ideas proposed for STR and end-to-end systems in the era of deep learning. The primary contribution of each approach is reviewed.  ... 
arXiv:2005.03492v3 fatcat:rmzmavxylnf6rbp52lje2mrgiy

Arabic Cursive Text Recognition from Natural Scene Images

Saad Ahmed, Saeeda Naz, Muhammad Razzak, Rubiyah Yusof
2019 Applied Sciences  
The presented techniques following a deep learning architecture are equally suitable for the development of Arabic cursive scene text recognition systems.  ...  Surveys on the Latin and Chinese script-based scene text recognition system can be found, but the Arabic like scene text recognition problem is yet to be addressed in detail.  ...  Acknowledgments: The authors would like to thank the Center for Artificial Intelligence and RObotics (CAIRO-ikhoza) lab under MJIIT, Universiti Teknologi Malaysia for arranging funds to conduct this research  ... 
doi:10.3390/app9020236 fatcat:jowekkffqrelvj3mk4324uj2v4

TextSR: Content-Aware Text Super-Resolution Guided by Recognition [article]

Wenjia Wang, Enze Xie, Peize Sun, Wenhai Wang, Lixun Tian, Chunhua Shen, Ping Luo
2019 arXiv   pre-print
Scene text recognition has witnessed rapid development with the advance of convolutional neural networks.  ...  Different from previous super-resolution methods, we use the loss of text recognition as the Text Perceptual Loss to guide the training of the super-resolution network, and thus it pays more attention  ...  In the deep learning era, super-resolution is simply treated as a regression problem, where the input is the low-resolution image, and the target output is the high-resolution image [5] .  ... 
arXiv:1909.07113v4 fatcat:vgpj3o7gcrbuxexuck2d2wdnuy

Optical Character Recognition using CRNN

Though there are many existing applications, we plan on exploring the domain of deep learning and build an optical character recognition system using deep learning architectures.  ...  Optical Character Recognition (OCR) is a computer vision technique which recognizes text present in any form of images, such as scanned documents and photos.  ...  The major phases include text detection and text recognition. The proposed system primarily focuses on the text recognition phase.  ... 
doi:10.35940/ijitee.h6264.069820 fatcat:x5znm7rkvvdsnnsl45fcrrjczi
« Previous Showing results 1 — 15 out of 15,950 results