Geometry Normalization Networks for Accurate Scene Text Detection

Hongbin Sun; Jiaqi Duan; Wayne Zhang; Xiaoyu Yue; Youjiang Xu; Yue Guan; Zhanghui Kuang

REVIEW 3 major objections 6 minor 43 references

Geometry Normalization Networks for Accurate Scene Text Detection

T0 review · 3 major / 6 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper claims that scene-text detectors fail on large geometry variance, and that a multi-branch Geometry Normalization Module that rescales and rotates features into canonical ranges lets one shared header reach 88.52 F-score on ICDAR…

desk verdict A solid, well-ablated architecture paper for scene text detection whose headline SOTA numbers rest on hand-set geometry ranges that the paper never checks cover the test annotations. read the letter →

arxiv 1909.00794 v1 pith:WR2CLVH3 submitted 2019-09-02 cs.CV

Youjiang Xu , Jiaqi Duan , Zhanghui Kuang , Xiaoyu Yue , Hongbin Sun , Yue Guan , Wayne Zhang This is my paper · ORCID

classification cs.CV

keywords scenetextdetectiongeometrynormalizationorientedscaleinvariancerotationconvolutionalneuralnetworksICDAR20152017MLT

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper argues that convolutional detectors for scene text fail on large geometry variance—text size and orientation—because a single detection header cannot learn all of it at once. To fix this, it introduces a Geometry Normalization Module with multiple branches, each rescaling and rotating or flipping features so that every text instance lands in a narrow canonical range through at least one branch. The branches feed one shared text-detection header, and a geometry-aware training scheme augments samples so that all branches are exercised uniformly. If the claim holds, plugging this module into existing detectors improves accuracy on oriented and multilingual benchmarks while keeping a single forward pass at test time.

What carries the argument

The load-bearing object is the Geometry Normalization Module (GNM), a set of parallel branches placed between the backbone feature extractor and a shared text-detection header. Each branch combines one Scale Normalization Unit (SNU)—either identity $S$ or downsample-by-half $S_{1/2}$—with one Orientation Normalization Unit (ONU): identity $O$, clockwise rotation $O_r$, horizontal flip $O_f$, or rotation-plus-flip $O_{r+f}$. These transforms map every text instance in a branch's feasible range into the same canonical geometry, so the header only ever sees a limited distribution; Table 2 gives the proxy ground-truth transforms that invert this normalization for training. The two-by-four branching keeps the canonical orientation interval at $[0,\pi/4]$ and the canonical scale range at roughly $[10,100]$ pixels.

What would settle it

Collect or synthesize a test set containing text instances deliberately outside the fixed branch ranges—for example, short sides below 10 pixels or angles near the endpoints of the feasible intervals—run the trained GNNets, and check whether those instances are missed by construction; if the ranges are load-bearing, recall on them will be near zero regardless of the learned features.

Watch

Extended reading notes

Core claim

The central claim is that large geometry variances can be normalized away before the detection header, instead of asking the header to absorb them. Experiments on a rotated version of ICDAR 2015 show that a strong detector such as EAST drops from 80.6 to 20.9 F-score under large orientation variance, and rotation augmentation alone cannot recover the loss. The proposed Geometry Normalization Module splits the geometry distribution into combinations of two scale units ($S$, $S_{1/2}$) and four orientation units ($O$, $O_r$, $O_f$, $O_{r+f}$), each branch transforming text in its feasible range into a canonical interval such as $[0,\pi/4]$ in angle and a compact scale range. A shared header then trains on proxy ground truths in canonical coordinates, and test predictions are back-projected and merged by NMS. With this design and the geometry-aware training strategy, GNNets reach one-forward F-scores of 88.52 on ICDAR 2015 and 74.54 on ICDAR 2017 MLT, surpassing prior state-of-the-art methods, including some that use multi-scale testing or extra recognition supervision.

Load-bearing premise

The feasible geometry ranges fixed for the normalization branches are assumed to cover every text instance that appears at test time, since the test strategy discards any predicted box outside its branch's range.

Editorial extensions

If this is right

Plugging the GNM into existing CNN detectors such as EAST and PSENet raises one-forward F-scores on rotated and multilingual benchmarks without multi-scale test-time fusion.
Because all branches share one detection header, the method avoids training separate per-geometry headers while still covering a wide geometry distribution.
The geometry-aware training strategy, which augments each sampled text instance seven times so every branch gets valid samples, appears necessary for the module to realize its gains.
The default configuration of two scale units and four orientation units is sufficient; adding further units yields only marginal improvements in the ablations.
The test-time rule that discards predictions outside a branch's feasible range means final accuracy depends on the hand-set geometry ranges covering the target distribution.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

A natural extension is to apply the same geometry-normalization idea to general object detection, where pose and aspect-ratio variance similarly burden a single header.
The fixed branch ranges could be learned or expanded during training, which would remove manual allocation and likely improve recall on long-tail text geometries.
Because normalization happens in feature space rather than by image pyramid, the module should combine well with faster backbones, offering accuracy gains at lower inference cost than multi-scale test methods.
The rotated-ICDAR evaluation suggests that standard benchmarks with mostly horizontal text understate orientation robustness; building rotated variants of other detection benchmarks would make geometry sensitivity visible.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Here is my read. The paper is a genuine and mostly honest architecture contribution. The GNM is simple: multi-branch feature map transformations that downscale and rotate/flip by fixed 90-degree steps, so text instances of varied scale and orientation land in a canonical range before a shared detection head. The proxy ground-truth table is deterministic and internally consistent. The geometry-aware training sampler (augmenting each instance seven times so every branch gets valid samples) is a practical addition. The ablation tables are systematic, and the gains on the self-constructed rotated ICDAR 2015 are large and believable: EAST jumps from 21 to 80.8 F-score. That is real evidence the module does what it claims.

The soft spot is the one the stress-test flags. Section 4.3 says test predictions outside a branch's feasible geometry range are discarded, and the feasible ranges are hand-set. The paper asserts the union of branch ranges equals the whole text geometry distribution, but never measures whether the actual test annotations fall inside that union after the resizing protocol. On ICDAR 2015 the union is [10,200] px in short-side length; any text shorter than 10 or longer than 200 after resizing to 2240 is zero-recall by construction. I suspect the fraction is tiny, but "suspect" is not a histogram. This is a genuine gap, not a manufactured one. Also worth noting: the orientation branches are evaluated only on the rotated set; the standard benchmark comparisons use scale-only branches, and the branch count and ranges are chosen by test-set performance. No error bars, no code.

None of this sinks the paper. The central idea is sound, the ablations support it, and the discussion of SNIP/SNIPER and RRD/ITN is fair. The coverage ceiling is a fixable validation gap, not a contradiction. I'd send this to peer review with a request for a coverage analysis on the actual benchmarks and ideally code release. It deserves referee time, and the questions are answerable in a revision.

Referee Report

3 major / 6 minor

Summary. The paper proposes a Geometry Normalization Module (GNM) that normalizes text instance geometry through multiple branches, each combining a Scale Normalization Unit (SNU) and an Orientation Normalization Unit (ONU), so that a shared text detection header sees a limited canonical geometry distribution. The authors also introduce a geometry-aware training scheme with augmentation, and instantiate GNNets on EAST and PSENet. On ICDAR 2015 and ICDAR 2017 MLT they report one-forward F-scores of 88.52 and 74.54, and on a rotated ICDAR 2015 benchmark they show large gains over baselines.

Significance. If the reported results hold, the paper makes a useful contribution by showing that explicit geometry normalization at the feature level (downsampling, rotation, flipping) can reduce the burden on a shared detection header, and by providing a new rotated benchmark. The ablation studies are systematic and support the central claim that narrowing the canonical geometry range improves accuracy. The module is architecture-agnostic and could be applied to other detectors. However, the headline claims are currently weakened by the unresolved coverage question and by test-set-based hyperparameter selection, so the practical significance is conditional on fixing these issues.

major comments (3)

[Section 4.3] The test-time discarding of predicted boxes that do not lie in a branch's feasible geometry range creates a hard coverage ceiling, yet the paper never verifies that the union of feasible ranges covers the actual test text geometry. In Section 4.3 the authors state that 'the union of the feasible geometry ranges of all branches equals to the whole text geometry distribution,' but no measurement or analysis is provided to support this for ICDAR 2015 or ICDAR 2017 MLT after the reported resizing protocol (long side 2240 and 3200, respectively). Since the GNM used on these benchmarks contains only SNU branches, any text instance with a short side below 10 px or above 200 px is undetectable by construction. Please report the empirical distribution of ground-truth short sides on these test sets after resizing, quantify the fraction falling outside [10,200], and either justify that the union is sufficient or acknowledge that the reported F-scores are upper bounds on the covered geometry subset.
[Section 5.3] The hyperparameters (number of SNU/ONU branches and the feasible/canonical geometry ranges) are selected using the final test sets: Table 3 uses the ICDAR 2015 test set to choose two SNU branches, and Table 4 uses the Rotated ICDAR 2015 test set to choose four ONU branches. The same ICDAR 2015 test set is then used for the headline comparison in Table 6, so the reported numbers are the result of test-set tuning and are likely optimistic. Please use a held-out validation set (e.g., ICDAR 2013 or a split of the training data) for model selection, or at minimum clearly state that the ICDAR 2015 result was selected on the test set and discuss the implications for the comparison.
[Table 6] The claimed improvements over strong baselines are small—e.g., +1.22 F-score over the reimplemented PSENet on ICDAR 2015 and +1.80 on ICDAR 2017 MLT—and no error bars or multiple-seed variation are reported. Given that the final configuration was selected on the test set, it is unclear whether these margins are statistically meaningful. Please report mean and standard deviation over at least three independent runs for the main models, or otherwise provide evidence of training stability, so the reader can judge whether the differences are significant.

minor comments (6)

[Abstract / Section 1] The F-score on ICDAR 2015 is reported as 88.52 in the abstract and 88.50 in Section 1; please unify the numbers.
[Section 5.3] The text says the reimplemented EAST obtains a Rotated ICDAR 2015 F-score of 21.99%, but Table 5(a) reports 20.99%; please correct the inconsistency.
[Table 4] The 'Range' column for the four-branch ONU is given as [0, π/4], but the feasible orientation range of the Or branch consists of two disjoint intervals ([−π/2,−π/4] and [π/4,π/2]); consider clarifying in the caption that 'Range' denotes the desired canonical orientation range, not the feasible range.
[Figure 3] The angle intervals in Figure 3 are difficult to read in the version provided; please ensure the figure is legible and that the caption clearly maps each ONU (O, Or, Of, Or+f) to its input and output angle ranges.
[Section 5.2] There is a typo: 'IDCAR 2015' should be 'ICDAR 2015'.
[General] The term 'one-forward test' is used frequently; please define it at first use for readers not familiar with the detection literature.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: GNM transformations are deterministic and the method is evaluated on external benchmarks.

full rationale

The paper's central mechanism is a deterministic geometry normalization: each GNM branch applies a fixed scale and orientation transform (Section 4.1), and the proxy ground truths in Table 2 are the exact inverse mappings of those transforms. The network is trained with a shared detection header on these transformed feature maps, and at test time predicted boxes are back-projected and merged via NMS. No parameter is fitted to the benchmark test set and then reported as a prediction: the reported F-scores come from the external ICDAR 2015 and ICDAR 2017 MLT test sets. The feasible geometry ranges are hand-set design choices, not fitted values, and the observation that the test-time filtering discards boxes outside those ranges is a coverage limitation rather than a circular step. The paper cites prior work, including a paper by overlapping authors, but the GNM architecture and training strategy are not justified by that citation; the ablation studies independently compare configurations. There is no equation here that reduces to its inputs by construction, and no self-citation is load-bearing. The skeptic's concern about hand-set ranges creating a coverage ceiling is a legitimate external-validity caveat, but it does not make the derivation circular. Therefore the appropriate circularity score is 0.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The central claim rests on hand-chosen geometry ranges and branch counts, an equivariance assumption for feature rotation, and a coverage assumption for the test geometry distribution. No invented physical entities are introduced.

free parameters (6)

canonical scale range = [10, 100] pixels
Target short-side range after two SNU branches; set by hand in Section 5.3 and validated by ablation (Table 3).
per-branch feasible scale ranges = S: [10, 80], S^{1/2}: [60, 200]
Hand-chosen overlapping intervals in Section 5.3; they define which branch is responsible for which text scale and determine the canonical range.
canonical orientation range = [0, pi/4]
Target angle range after four ONU branches; fixed by the four-quadrant partition in Section 4.1 and validated in Table 4.
number of SNU branches = 2
Selected in Section 5.3 because adding a third branch (S^{1/4}) reduced F-score on ICDAR 2015.
number of ONU branches = 4
Selected in Section 5.3 because the four-branch orientation set gave the best F-score on Rotated ICDAR 2015.
number of augmented samples per text instance = 7
Hand-chosen in Section 4.3 training sampling strategy so every branch receives valid instances each batch.

assumptions (4)

domain assumption The CNN feature extractor and the 3x3 conv layers in the normalization units are (approximately) equivariant to 90-degree rotation and horizontal flip, so rotating or flipping feature maps is equivalent to transforming the input image geometry.
Section 4.1 defines ONU branches that rotate/flip feature maps; exact equivariance holds only for aligned discrete rotations and not for arbitrary angles, yet the method relies on it to bring text into a canonical angle range.
domain assumption The union of feasible geometry ranges over branches covers the entire test geometry distribution, so no test text is discarded by the test strategy.
Section 4.3 states the union equals the whole text geometry distribution; Section 4.3 test strategy discards outputs outside a branch's feasible range. If test text falls outside the chosen ranges, detection fails by construction.
domain assumption The reimplemented EAST and PSENet baselines match the original implementations closely enough to serve as fair baselines and initialization points.
Section 5.4 says differences from original PSENet are below 0.4%; the ablation tables and final comparisons rely on these reimplementations.
domain assumption The original EAST/PSENet loss functions remain valid when applied to the proxy ground truth coordinates transformed by Table 2.
Section 4.2 Eq. (4) reuses the detector losses on transformed boxes; if the transformation broke the loss's geometric meaning, training would be impaired.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Geometry Normalization Networks for Accurate Scene Text Detection." pith.science (2026). https://pith.science/paper/WR2CLVH3

@misc{pith2026190900794,
  author       = {Pith},
  title        = {Pith review of: Geometry Normalization Networks for Accurate Scene Text Detection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WR2CLVH3}},
  note         = {Machine review of arXiv:1909.00794}
}

read the original abstract

Large geometry (e.g., orientation) variances are the key challenges in the scene text detection. In this work, we first conduct experiments to investigate the capacity of networks for learning geometry variances on detecting scene texts, and find that networks can handle only limited text geometry variances. Then, we put forward a novel Geometry Normalization Module (GNM) with multiple branches, each of which is composed of one Scale Normalization Unit and one Orientation Normalization Unit, to normalize each text instance to one desired canonical geometry range through at least one branch. The GNM is general and readily plugged into existing convolutional neural network based text detectors to construct end-to-end Geometry Normalization Networks (GNNets). Moreover, we propose a geometry-aware training scheme to effectively train the GNNets by sampling and augmenting text instances from a uniform geometry variance distribution. Finally, experiments on popular benchmarks of ICDAR 2015 and ICDAR 2017 MLT validate that our method outperforms all the state-of-the-art approaches remarkably by obtaining one-forward test F-scores of 88.52 and 74.54 respectively.

Figures

Figures reproduced from arXiv: 1909.00794 by the authors.

**Figure 2.** The framework of the proposed Geometry Normalization Networks. The feature maps extracted by the backbone are fed into the [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗

**Figure 3.** The changes of the text box orientation by applying the [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

Figures from the paper (1 more)

**Figure 4.** Figure 4: Qualitative results on ICDAR 2015, ICDAR 2017 MLT and Rotated ICDAR 2015. We compare the results generated by PSENet [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 40 canonical work pages

[1]

http://rrc.cvc

ICDAR2017 Competition on Multi-Lingual Scene Text De- tection and Script Identiﬁcation. http://rrc.cvc. uab.es/?ch=8&com=introduction, 2017

work page 2017
[2]

Fastext: Ef- ﬁcient Unconstrained Scene Text Detector

Michal Busta, Lukas Neumann, and Jiri Matas. Fastext: Ef- ﬁcient Unconstrained Scene Text Detector. In ICCV, 2015

work page 2015
[3]

Fused Text Segmentation Networks for Multi-Oriented Scene Text Detection

Yuchen Dai, Zheng Huang, Yuting Gao, Youxuan Xu, Kai Chen, Jie Guo, and Weidong Qiu. Fused Text Segmentation Networks for Multi-Oriented Scene Text Detection. InICPR, 2018

work page 2018
[4]

Detecting Text in Natural Scenes with Stroke Width Transform

Boris Epshtein, Eyal Ofek, and Yonatan Wexler. Detecting Text in Natural Scenes with Stroke Width Transform. In CVPR, 2010

work page 2010
[5]

Fast R-CNN

Ross Girshick. Fast R-CNN. In ICCV, 2015

work page 2015
[6]

Rich Feature Hierarchies for Accurate Object Detec- tion and Semantic Segmentation

Ross Girshick, Jeff Donahue, Trevor Darrell, and Jitendra Malik. Rich Feature Hierarchies for Accurate Object Detec- tion and Semantic Segmentation. In CVPR, 2014

work page 2014
[7]

Single Shot Text Detector with Regional Attention

Pan He, Weilin Huang, Tong He, Qile Zhu, Yu Qiao, and Xi- aolin Li. Single Shot Text Detector with Regional Attention. In ICCV, 2017

work page 2017
[8]

WordSup: Exploiting Word Annotations for Character Based Text Detection

Han Hu, Chengquan Zhang, Yuxuan Luo, Yuzhuo Wang, Junyu Han, and Errui Ding. WordSup: Exploiting Word Annotations for Character Based Text Detection. In ICCV, 2017

work page 2017

Show all 43 references

[9]

Text Localization in Natural Images using Stroke Feature Trans- form and Text Covariance Descriptors

Weilin Huang, Zhe Lin, Jianchao Yang, and Jue Wang. Text Localization in Natural Images using Stroke Feature Trans- form and Text Covariance Descriptors. In ICCV, 2013

2013
[10]

R2cnn: Rota- tional region cnn for orientation robust scene text detection

Yingying Jiang, Xiangyu Zhu, Xiaobing Wang, Shuli Yang, Wei Li, Hua Wang, Pei Fu, and Zhenbo Luo. R2cnn: Rota- tional region cnn for orientation robust scene text detection. arXiv preprint arXiv:1706.09579, 2017

2017 arXiv
[11]

ICDAR 2015 Competition on Robust Reading

Dimosthenis Karatzas, Lluis Gomez-Bigorda, Anguelos Nicolaou, Suman Ghosh, Andrew Bagdanov, Masakazu Iwa- mura, Jiri Matas, Lukas Neumann, Vijay Ramaseshan Chan- drasekhar, Shijian Lu, and Others. ICDAR 2015 Competition on Robust Reading. In ICDAR, 2015

2015
[12]

Icdar 2013 robust read- ing competition

Dimosthenis Karatzas, Faisal Shafait, Seiichi Uchida, Masakazu Iwamura, Lluis Gomez i Bigorda, Sergi Robles Mestre, Joan Mas, David Fernandez Mota, Jon Almazan Al- mazan, and Lluis Pere De Las Heras. Icdar 2013 robust read- ing competition. In ICDAR, 2013

2013
[13]

Adam: A method for stochastic optimization

Diederik P Kingma and Jimmy Ba. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980 , 2014

2014 arXiv
[14]

Shape Robust Text Detection with Progressive Scale Expansion Network

Xiang Li, Wenhai Wang, Wenbo Hou, Ruo-Ze Liu, Tong Lu, and Jian Yang. Shape Robust Text Detection with Progressive Scale Expansion Network. arXiv preprint arXiv:1806.02559, 2018

2018 arXiv
[15]

Pixel-anchor: A fast oriented scene text detector with combined networks

Yuan Li, Yuanjie Yu, Zefeng Li, Yangkun Lin, Meifang Xu, Jiwei Li, and Xi Zhou. Pixel-anchor: A fast oriented scene text detector with combined networks. arXiv preprint arXiv:1811.07432, 2018

2018 arXiv
[16]

TextBoxes: a Fast Text Detector with a Single Deep Neural Network

Minghui Liao, Baoguang Shi, Xiang Bai, Xinggang Wang, and Wenyu Liu. TextBoxes: a Fast Text Detector with a Single Deep Neural Network. In AAAI, 2017

2017
[17]

Rotation-Sensitive Regression for Oriented Scene Text Detection

Minghui Liao, Zhen Zhu, Baoguang Shi, Gui-song Xia, and Xiang Bai. Rotation-Sensitive Regression for Oriented Scene Text Detection. In CVPR, 2018

2018
[18]

Feature Pyramid Networks for Object Detection

Tsung-Yi Lin, Piotr Doll ´ar, Ross Girshick, Kaiming He, Bharath Hariharan, and Serge Belongie. Feature Pyramid Networks for Object Detection. In CVPR, 2017

2017
[19]

Wei Liu, Dragomir Anguelov, Dumitru Erhan, Christian Szegedy, Scott Reed, Cheng Yang Fu, and Alexander C. Berg. SSD: Single Shot Multibox Detector. In ECCV, 2016

2016
[20]

Fots: Fast Oriented Text Spotting with a Uniﬁed Network

Xuebo Liu, Ding Liang, Shi Yan, Dagui Chen, Yu Qiao, and Junjie Yan. Fots: Fast Oriented Text Spotting with a Uniﬁed Network. In CVPR, 2018

2018
[21]

Deep Matching Prior Net- work: Toward Tighter Multi-oriented Text Detection

Yuliang Liu and Lianwen Jin. Deep Matching Prior Net- work: Toward Tighter Multi-oriented Text Detection. In CVPR, 2017

2017
[22]

Fully Convolutional Networks for Semantic Segmentation

Jonathan Long, Evan Shelhamer, and Trevor Darrell. Fully Convolutional Networks for Semantic Segmentation. In CVPR, 2015

2015
[23]

TextSnake : A Flexible Representation for Arbitrary Shapes

Shangbang Long, Jiaqiang Ruan, Wenjie Zhang, and Xin He. TextSnake : A Flexible Representation for Arbitrary Shapes. In ECCV, 2018

2018
[24]

Mask TextSpotter: An End-to-End Trainable Neural Network for Spotting Text with Arbitrary Shapes

Pengyuan Lyu, Minghui Liao, Cong Yao, Wenhao Wu, and Xiang Bai. Mask TextSpotter: An End-to-End Trainable Neural Network for Spotting Text with Arbitrary Shapes. In ECCV, 2018

2018
[25]

Multi-Oriented Scene Text Detection via Corner Localization and Region Segmentation

Pengyuan Lyu, Cong Yao, Wenhao Wu, Shuicheng Yan, and Xiang Bai. Multi-Oriented Scene Text Detection via Corner Localization and Region Segmentation. In CVPR, 2018

2018
[26]

IC- DAR2017 Robust Reading Challenge on Multi-Lingual Scene Text Detection and Script Identiﬁcation-RRC-MLT

Nibal Nayef, Fei Yin, Imen Bizid, Hyunsoo Choi, Yuan Feng, Dimosthenis Karatzas, Zhenbo Luo, Umapada Pal, Christophe Rigaud, Joseph Chazalon, and Others. IC- DAR2017 Robust Reading Challenge on Multi-Lingual Scene Text Detection and Script Identiﬁcation-RRC-MLT. In ICDAR, 2017

2017
[27]

A Method for Text Local- ization and Recognition in Real-World Images

Lukas Neumann and Jiri Matas. A Method for Text Local- ization and Recognition in Real-World Images. In ACCV, 2010

2010
[28]

Real-time Scene Text Lo- calization and Recognition

Lukas Neumann and Jiri Matas. Real-time Scene Text Lo- calization and Recognition. In CVPR, 2012

2012
[29]

You Only Look Once: Uniﬁed, Real-Time Object Detection

Joseph Redmon, Santosh Divvala, Ross Girshick, and Ali Farhadi. You Only Look Once: Uniﬁed, Real-Time Object Detection. In CVPR, 2016

2016
[30]

Faster RCNN: Towards Real-time Object Detection with Re- gion Proposal Networks

Shaoqing Ren, Kaiming He, Ross Girshick, and Jian Sun. Faster RCNN: Towards Real-time Object Detection with Re- gion Proposal Networks. In NIPS, 2015

2015
[31]

Detecting Oriented Text in Natural Images by Linking Segments

Baoguang Shi, Xiang Bai, and Serge Belongie. Detecting Oriented Text in Natural Images by Linking Segments. In CVPR, 2017

2017
[32]

An Analysis of Scale In- variance in Object Detection SNIP

Bharat Singh and Larry S Davis. An Analysis of Scale In- variance in Object Detection SNIP. In CVPR, 2018

2018
[33]

SNIPER: Efﬁcient multi-scale training

Bharat Singh, Mahyar Najibi, and Larry S Davis. SNIPER: Efﬁcient multi-scale training. In NIPS, 2018

2018
[34]

De- tecting Text in Natural Image with Connectionist Text Pro- posal Network

Zhi Tian, Weilin Huang, Tong He, Pan He, and Yu Qiao. De- tecting Text in Natural Image with Connectionist Text Pro- posal Network. In ECCV, 2016

2016
[35]

Text Localization and Recognition in Images and Video

Seiichi Uchida. Text Localization and Recognition in Images and Video. Handbook of Document Image Processing and Recognition, 2014

2014
[36]

Geometry-Aware Scene Text Detection with Instance Transformation Network

Fangfang Wang, Liming Zhao, Xi Li, Xinchao Wang, and Dacheng Tao. Geometry-Aware Scene Text Detection with Instance Transformation Network. In CVPR, 2018

2018
[37]

Text Detection and Recognition in Imagery: A Survey

Qixiang Ye and David Doermann. Text Detection and Recognition in Imagery: A Survey. PAMI, 2015

2015
[38]

Multi-orientation Scene Text Detection with Adaptive Clus- tering

Xu-Cheng Yin, Wei-Yi Pei, Jun Zhang, and Hong-Wei Hao. Multi-orientation Scene Text Detection with Adaptive Clus- tering. PAMI, 2015

2015
[39]

Boosting up Scene Text Detectors with Guided CNN

Xiaoyu Yue, Zhanghui Kuang, Zhaoyang Zhang, Zhenfang Chen, Pan He, Yu Qiao, and Wayne Zhang. Boosting up Scene Text Detectors with Guided CNN. In BMVC, 2018

2018
[40]

Symmetry-based Text Line Detection in Natural Scenes

Zheng Zhang, Wei Shen, Cong Yao, and Xiang Bai. Symmetry-based Text Line Detection in Natural Scenes. In CVPR, 2015

2015
[41]

Deeptext: A Uniﬁed Framework for Text Proposal Generation and Text Detection in Natural Images

Zhuoyao Zhong, Lianwen Jin, Shuye Zhang, and Ziyong Feng. Deeptext: A Uniﬁed Framework for Text Proposal Generation and Text Detection in Natural Images. arXiv preprint arXiv:1605.07314, 2016

2016 arXiv
[42]

EAST: An Efﬁcient and Accurate Scene Text Detector

Xinyu Zhou, Cong Yao, He Wen, Yuzhi Wang, Shuchang Zhou, Weiran He, and Jiajun Liang. EAST: An Efﬁcient and Accurate Scene Text Detector. In CVPR, 2017

2017
[43]

Scene Text Detec- tion and Recognition: Recent Advances and Future Trends

Yingying Zhu, Cong Yao, and Xiang Bai. Scene Text Detec- tion and Recognition: Recent Advances and Future Trends. Frontiers of Computer Science, 2016

2016

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

http://rrc.cvc

ICDAR2017 Competition on Multi-Lingual Scene Text De- tection and Script Identiﬁcation. http://rrc.cvc. uab.es/?ch=8&com=introduction, 2017

work page 2017

[2] [2]

Fastext: Ef- ﬁcient Unconstrained Scene Text Detector

Michal Busta, Lukas Neumann, and Jiri Matas. Fastext: Ef- ﬁcient Unconstrained Scene Text Detector. In ICCV, 2015

work page 2015

[3] [3]

Fused Text Segmentation Networks for Multi-Oriented Scene Text Detection

Yuchen Dai, Zheng Huang, Yuting Gao, Youxuan Xu, Kai Chen, Jie Guo, and Weidong Qiu. Fused Text Segmentation Networks for Multi-Oriented Scene Text Detection. InICPR, 2018

work page 2018

[4] [4]

Detecting Text in Natural Scenes with Stroke Width Transform

Boris Epshtein, Eyal Ofek, and Yonatan Wexler. Detecting Text in Natural Scenes with Stroke Width Transform. In CVPR, 2010

work page 2010

[5] [5]

Fast R-CNN

Ross Girshick. Fast R-CNN. In ICCV, 2015

work page 2015

[6] [6]

Rich Feature Hierarchies for Accurate Object Detec- tion and Semantic Segmentation

Ross Girshick, Jeff Donahue, Trevor Darrell, and Jitendra Malik. Rich Feature Hierarchies for Accurate Object Detec- tion and Semantic Segmentation. In CVPR, 2014

work page 2014

[7] [7]

Single Shot Text Detector with Regional Attention

Pan He, Weilin Huang, Tong He, Qile Zhu, Yu Qiao, and Xi- aolin Li. Single Shot Text Detector with Regional Attention. In ICCV, 2017

work page 2017

[8] [8]

WordSup: Exploiting Word Annotations for Character Based Text Detection

Han Hu, Chengquan Zhang, Yuxuan Luo, Yuzhuo Wang, Junyu Han, and Errui Ding. WordSup: Exploiting Word Annotations for Character Based Text Detection. In ICCV, 2017

work page 2017