Filters








30,705 Hits in 6.4 sec

Learning to Predict More Accurate Text Instances for Scene Text Detection [article]

XiaoQian Li, Jie Liu, ShuWu Zhang, GuiXuan Zhang
2020 arXiv   pre-print
Furthermore, to predict more accurate text instances, the text instance accuracy loss is proposed as an assistant task to refine the predicted coordinates under the guidance of IoU.  ...  Nevertheless, there are still some difficulties for arbitrary shape text detection, especially for a simple and proper representation of arbitrary shape text instances.  ...  text instance accuracy loss of more accurate polygon prediction task.  ... 
arXiv:1911.07423v2 fatcat:5ecnsizt3ja5lhne7uyzessjz4

TextContourNet: a Flexible and Effective Framework for Improving Scene Text Detection Architecture with a Multi-task Cascade [article]

Dafang He, Xiao Yang, Daniel Kifer, C.Lee Giles
2018 arXiv   pre-print
We study the problem of extracting text instance contour information from images and use it to assist scene text detection.  ...  We propose two ways for learning the contour task together with the scene text detection: (1) as an auxiliary task and (2) as multi-task cascade.  ...  Being able to provide instance-level information in the early stage of the network allows the network to learn to propose an instance bounding box more easily and accurately.  ... 
arXiv:1809.03050v2 fatcat:3itocgvgrfherj3v7f5bdy5qoi

TextField: Learning A Deep Direction Field for Irregular Scene Text Detection [article]

Yongchao Xu, Yukang Wang, Wei Zhou, Yongpan Wang, Zhibo Yang, Xiang Bai
2019 arXiv   accepted
It is of great interest to detect curved texts, which are actually very common in natural scenes. In this paper, we present a novel text detector named TextField for detecting irregular scene texts.  ...  It encodes both binary text mask and direction information used to separate adjacent text instances, which is challenging for classical segmentation-based approaches.  ...  Scene text detection Scene text detection methods can be roughly classified into specifically engineered and deep learning-based methods.  ... 
arXiv:1812.01393v2 fatcat:npvethdp3bdmtck6bvmfx4mddi

MSR: Multi-Scale Shape Regression for Scene Text Detection [article]

Chuhui Xue, Shijian Lu, Wei Zhang
2019 arXiv   pre-print
The proposed MSR detects scene texts by predicting dense text boundary points that inherently capture the location and shape of text lines accurately and are also more tolerant to the variation of text  ...  State-of-the-art scene text detection techniques predict quadrilateral boxes that are prone to localization errors while dealing with straight or curved text lines of different orientations and lengths  ...  The large scale variation often leads to miss detection for ultra-small text instances or broken detection for ultra-large text instances.  ... 
arXiv:1901.02596v2 fatcat:dfijqa75jjccre4vrrzzv2q27q

Mask is All You Need: Rethinking Mask R-CNN for Dense and Arbitrary-Shaped Scene Text Detection [article]

Xugong Qin, Yu Zhou, Youhui Guo, Dayan Wu, Zhihong Tian, Ning Jiang, Hongbin Wang, Weiping Wang
2021 arXiv   pre-print
Due to the large success in object detection and instance segmentation, Mask R-CNN attracts great attention and is widely adopted as a strong baseline for arbitrary-shaped scene text detection and spotting  ...  And we propose instance-aware mask learning in which the mask head learns to predict the shape of the whole instance rather than classify each pixel to text or non-text.  ...  RELATED WORK According to different perspectives for modeling scene text, scene text detection can be roughly divided into bottom-up and top-down methods.  ... 
arXiv:2109.03426v1 fatcat:oc7is6an6rdexmgwuyigea7bde

Which and Where to Focus: A Simple yet Accurate Framework for Arbitrary-Shaped Nearby Text Detection in Scene Images [article]

Youhui Guo, Yu Zhou, Xugong Qin, Weiping Wang
2021 arXiv   pre-print
In particular, confusion problem arises in the case of nearby text instances. In this paper, we propose a simple yet effective method for accurate arbitrary-shaped nearby scene text detection.  ...  Firstly, a One-to-Many Training Scheme (OMTS) is designed to eliminate confusion and enable the proposals to learn more appropriate groundtruths in the case of nearby text instances.  ...  For each axis-aligned box proposal, with a more appropriate feature extracted by PFAM for arbitraryshaped text instances, we can obtain more accurate detection results.  ... 
arXiv:2109.03451v1 fatcat:mwrofrbbjzgmdguzfjdul5qfqu

Pyramid Mask Text Detector [article]

Jingchao Liu, Xuebo Liu, Jie Sheng, Ding Liang, Xin Li, Qingjie Liu
2019 arXiv   pre-print
Scene text detection, an essential step of scene text recognition system, is to locate text instances in natural scene images automatically.  ...  text mask for each text instance.  ...  labeling for the instance boundary. • We introduce a novel plane clustering algorithm to find better text box with the 3D coordinate, which predicts more accurate text box and improves the robustness  ... 
arXiv:1903.11800v1 fatcat:zleabbqd3fcznhsok5dzx5tosi

FC2RN: A Fully Convolutional Corner Refinement Network for Accurate Multi-Oriented Scene Text Detection [article]

Xugong Qin, Yu Zhou, Dayan Wu, Yinliang Yue, Weiping Wang
2020 arXiv   pre-print
Recent scene text detection works mainly focus on curve text detection. However, in real applications, the curve texts are more scarce than the multi-oriented ones.  ...  imperfect detections, especially for long texts due to the limitation of the receptive field.  ...  IoU threshold of 0.5 is usually adopted in detection. However, it is not enough for accurate scene text detection and subsequent text recognition task.  ... 
arXiv:2007.05113v1 fatcat:yqnbacvdvreppnitks2gmr32je

Mask TextSpotter v3: Segmentation Proposal Network for Robust Scene Text Spotting [article]

Minghui Liao, Guan Pang, Jing Huang, Tal Hassner, Xiang Bai
2020 arXiv   pre-print
Recent end-to-end trainable methods for scene text spotting, integrating detection and recognition, showed much progress.  ...  Furthermore, the accurate proposals produced by SPN allow masked RoI features to be used for decoupling neighboring text instances.  ...  By comparison, the proposals of our SPN are more accurate, thereby producing only a single text instance for each RoI feature and leading to accurate detection/recognition results.  ... 
arXiv:2007.09482v1 fatcat:myfhwzw2crh3jev27jcimij4du

Bidirectional Regression for Arbitrary-Shaped Text Detection [article]

Tao Sheng, Zhouhui Lian
2021 arXiv   pre-print
Besides, a corresponding post-processing algorithm is also designed to sequentially combine the four prediction results and reconstruct the text instance accurately.  ...  Arbitrary-shaped text detection has recently attracted increasing interests and witnessed rapid development with the popularity of deep learning algorithms.  ...  be problematic for detecting scene texts.  ... 
arXiv:2107.06129v1 fatcat:otg5ktymczfgrcc5pclzr2nymq

Challenges of Deep Learning-based Text Detection in the Wild

Zobeir Raisi, Mohamed A. Naiel, Paul Fieguth, Steven Wardell, John Zelek
2021 Journal of Computational Vision and Imaging Systems  
The reported accuracy of recent state-of-the-art text detection methods, mostly deep learning approaches, is in the order of 80% to 90% on standard benchmark datasets.  ...  The objective of the paper is to quantify the current shortcomings and to identify the challenges for future text detection research.  ...  , ON, Canada for supporting this research work.  ... 
doi:10.15353/jcvis.v6i1.3543 fatcat:27j25weqmveabmf7vrovpt62je

Multi-scale FCN with Cascaded Instance Aware Segmentation for Arbitrary Oriented Word Spotting in the Wild

Dafang He, Xiao Yang, Chen Liang, Zihan Zhou, Alexander G. Ororbia, Daniel Kifer, C. Lee Giles
2017 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
Scene text detection has attracted great attention these years. Text potentially exist in a wide variety of images or videos and play an important role in understanding the scene.  ...  (2) a novel instance (word or line) aware segmentation is designed to further remove false positives and obtain word instances.  ...  By joint prediction, it can capture larger context information and give more accurate prediction.  ... 
doi:10.1109/cvpr.2017.58 dblp:conf/cvpr/HeYLZOKG17 fatcat:f3zydqfcdvc4rn4xq7drwl7kgm

CentripetalText: An Efficient Text Instance Representation for Scene Text Detection [article]

Tao Sheng, Jie Chen, Zhouhui Lian
2022 arXiv   pre-print
Scene text detection remains a grand challenge due to the variation in text curvatures, orientations, and aspect ratios.  ...  For the task of end-to-end scene text recognition, our method outperforms Mask TextSpotter v3 by 1.1% on Total-Text.  ...  How to represent text instances in real imagery is one of the major challenges for scene text detection, and usually there are two strategies to solve the problem arising from this challenge.  ... 
arXiv:2107.05945v3 fatcat:yvcbjajvbrfhnapyldqgobqchm

A robust and effective text detector supervised by Contrastive Learning

Ran Wei, Yaoyi Li, Haiyan Li, Ze Tang, Hongtao Lu, Nengbin Cai, Xuejun Zhao
2021 IEEE Access  
So far, the detection results for text instances in motion blur, low-resolution images are still not satisfactory.  ...  INDEX TERMS Scene text detection, contrastive learning, data augmentation.  ...  TEXT DETECTION For a long period of time, scene text detection and recognition in natural scenes have been popular research topics in computer vision.  ... 
doi:10.1109/access.2021.3057108 fatcat:apcqjj76prh3bh3dfmr2a326n4

Location-Aware Feature Selection Text Detection Network [article]

Zengyuan Guo, Zilin Wang, Zhihui Wang, Wanli Ouyang, Haojie Li, Wen Gao
2020 arXiv   pre-print
As a result, LASNet predicts the more accurate bounding boxes by using a learnable feature selection way.  ...  To address this issue, we propose a novel Location-Aware feature Selection text detection Network (LASNet).  ...  Then, we design a Location-Aware Feature Selection (LAFS) mechanism to select the components with the highest confidence score to form a more accurate bounding box for a text instance.  ... 
arXiv:2004.10999v2 fatcat:27rnkktwzfbc7cbxuq5v4lljr4
« Previous Showing results 1 — 15 out of 30,705 results