PAN++: Towards Efficient and Accurate End-to-End Spotting of Arbitrarily-Shaped Text
The DOI matches its Crossref record; 77 of the 77 references it cites were found in a registry or an index.
Abstract
Scene text detection and recognition have been well explored in the past few years. Despite the progress, efficient and accurate end-to-end spotting of arbitrarily-shaped text remains challenging. In this work, we propose an end-to-end text spotting framework, termed PAN++, which can efficiently detect and recognize text of arbitrary shapes in natural scenes. PAN++ is based on the kernel representation that reformulates a text line as a text kernel (central region) surrounded by peripheral pixels. By systematically comparing with existing scene text representations, we show that our kernel representation can not only describe arbitrarily-shaped text but also well distinguish adjacent text. Moreover, as a pixel-based representation, the kernel representation can be predicted by a single fully convolutional network, which is very friendly to real-time applications. Taking the advantages of the kernel representation, we design a series of components as follows: 1) a computationally efficient feature enhancement network composed of stacked Feature Pyramid Enhancement Modules (FPEMs); 2) a lightweight detection head cooperating with Pixel Aggregation (PA); and 3) an efficient attention-based recognition head with Masked RoI. Benefiting from the kernel representation and the tailored components, our method achieves high inference speed while maintaining competitive accuracy. Extensive experiments show the superiority of our method. For example, the proposed PAN++ achieves an end-to-end text spotting F-measure of 64.9 at 29.2 FPS on the Total-Text dataset, which significantly outperforms the previous best method. Code will be available at: git.io/PAN.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474 · cited 13×
“In 13 these experiments, we use ResNet18 as the default backbone network.”
MethodInfluentialVaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L., & Polosukhin, I. (2017) · arXiv · doi:10.48550/arxiv.1706.03762 · Cited by 120,404
Very Deep Convolutional Networks for Large-Scale Image Recognition
Simonyan, K., & Zisserman, A. (2014) · arXiv · doi:10.48550/arxiv.1409.1556 · Cited by 114,148 · cited 1×
“In virtue of the powerful representation of deep neural networks [1], [2], scene text detection and recognition has witnessed great progress in the past a few years [3], [4], [5], [6], [7], [8].”
BackgroundHochreiter, S., & Schmidhuber, J. (1997) · Neural Computation · doi:10.1162/neco.1997.9.8.1735 · Cited by 102,318 · cited 12×
“At time step 0 (initial time step), the decoder takes SOS’s feature vector fs and zero LSTM initial states as input.”
BackgroundInfluentialAdam: A Method for Stochastic Optimization
Kingma, D. P., & Ba, J. (2014) · arXiv · doi:10.48550/arxiv.1412.6980 · Cited by 98,946 · cited 1×
“All models are optimized by using ADAM [66] optimizer with a batch size of 16 on 4 GPUs.”
MethodFaster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
Ren, S., He, K., Girshick, R., & Sun, J. (2017) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2016.2577031 · Cited by 73,472 · cited 4×
“3(f), where an attention-based text recognition head was added on Mask R-CNN to simultaneously detect and recognize irregular text lines.”
MethodInfluentialImageNet: A large-scale hierarchical image database
Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., & Fei-Fei, L. (2009) · 2009 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2009.5206848 · Cited by 65,333 · cited 1×
“In these experiments, we use the ResNet [2] pre-trained on ImageNet [65] as the backbone network of our method.”
MethodGradient-based learning applied to document recognition
Lecun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998) · Proceedings of the IEEE · doi:10.1109/5.726791 · Cited by 58,805 · cited 2×
““Conv” and “BN” represent regular convolution [49] and Batch Normalization [50], respectively.”
BackgroundFully convolutional networks for semantic segmentation
Long, J., Shelhamer, E., & Darrell, T. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7298965 · Cited by 38,834
Feature Pyramid Networks for Object Detection
Lin, T.-Y., Dollar, P., Girshick, R., He, K., Hariharan, B., & Belongie, S. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.106 · Cited by 30,939 · cited 8×
“Second, we compare our feature enhancement network (2- stacked FPEMs) with other widely used methods, including FPN [51], PAFPN [74], and NAS-FPN [75].”
BackgroundMethodInfluential
Research works printed but not linked: 1
- [58] C. Yao, X. Bai, W. Liu, Y. Ma, and Z. Tu, 'Detecting texts of arbitrary orientations in natural images,' in Proc. IEEE Conf. Comp. Vis. Patt. Recogn. , 2012. - · DOI · Paper page
Cited by
Pyramid Vision Transformer: A Versatile Backbone for Dense Prediction without Convolutions
Wang, W., Xie, E., Li, X., Fan, D.-P., Song, K., Liang, D., Lu, T., Luo, P., & Shao, L. (2021) · 2021 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.00061 · Cited by 5,259
Self-citationMPDIoU: A Loss for Efficient and Accurate Bounding Box RegressionPreprint
Ma, S., & Xu, Y. D. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2307.07662 · Cited by 832
On the Arbitrary-Oriented Object Detection: Classification Based Approaches Revisited
Yang, X., & Yan, J. (2022) · International Journal of Computer Vision · doi:10.1007/s11263-022-01593-w · Cited by 195
“SBD [20], SegLink [89], TextField [62], DBNet [33], PAN++ [64], PolarMask++ [65]), model ensemble (e.”
BackgroundSwinTextSpotter: Scene Text Spotting via Better Synergy between Text Detection and Text Recognition
Huang, M., Liu, Y., Peng, Z., Liu, C., Lin, D., Zhu, S., Yuan, N., Ding, K., & Jin, L. (2022) · 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52688.2022.00455 · Cited by 167
DeepSolo: Let Transformer Decoder with Explicit Points Solo for Text Spotting
Ye, M., Zhang, J., Zhao, S., Liu, J., Liu, T., Du, B., & Tao, D. (2023) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52729.2023.01854 · Cited by 142
Adapting a Swin Transformer for License Plate Number and Text Detection in Drone Images
Pal, S., Roy, A., Shivakumara, P., & Pal, U. (2023) · Artificial Intelligence and Applications · doi:10.47852/bonviewaia3202549 · Cited by 115
ABINet++: Autonomous, Bidirectional and Iterative Language Modeling for Scene Text Spotting
Fang, S., Mao, Z., Xie, H., Wang, Y., Yan, C. C., & Zhang, Y. (2022) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2022.3223908 · Cited by 107
SPTS: Single-Point Text Spotting
Peng, D., Wang, X., Liu, Y., Zhang, J., Huang, M., Lai, S., Li, J., Zhu, S., Lin, D., Shen, C., et al. (2022) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3503161.3547942 · Cited by 97
“, segmentation-based [18, 31, 32, 39] and regression-based methods [8, 22, 37].”
Self-citationOMNIPARSER: A Unified Framework for Text Spotting, Key Information Extraction and Table Recognition
Wan, J., Song, S., Yu, W., Liu, Y., Cheng, W., Huang, F., Bai, X., Yao, C., & Yang, Z. (2024) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52733.2024.01481 · Cited by 134
ESTextSpotter: Towards Better Scene Text Spotting with Explicit Synergy in Transformer
Huang, M., Zhang, J., Peng, D., Lu, H. F., Huang, C., Liu, Y., Bai, X., & Jin, L. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.01786 · Cited by 68
Yang, C., Chen, M., Yuan, Y., & Wang, Q. (2023) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2023.3289327 · Cited by 48
“Wang et al. [8], [39] proposed a faster region extension strategy based on their previous work [40], which could effectively fit arbitrary-shaped text instances and overcome the text adhesion problem.”
BackgroundYou Can even Annotate Text with Voice: Transcription-only-Supervised Text Spotting
Tang, J., Qiao, S., Cui, B., Ma, Y., Zhang, S., & Kanoulas, D. (2022) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3503161.3547787 · Cited by 86
“…methods CharNet [45] 66.2 - ABCNet [26] 64.2 75.7 PGNet [42] 63.1 - Mask TextSpotter [29] 65.3 77.4 Qin et al [34] 67.8 - Mask TextSpotter v3 [22] 71.2 78.4 MANGO [32] 72.9 83.6 PAN++ [43] 68.6 78.6 ABCNet v2 [27] 70.4 78.1 Point-based methods SPTS (Single-Point) [31] 67.9 74. directions.”
InfluentialXue, C., Zhang, W., Yu, H., Lu, S., Torr, P. H. S., & Bai, S. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-19815-1_17 · Cited by 64
Yu, W., Yang, Z., Wan, J., Song, S., Tang, J., Cheng, W., Liu, Y., & Bai, X. (2026) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2026.3677075 · Cited by 47
Assistive systems for visually impaired people: A survey on current requirements and advancements
Kathiria, P., Mankad, S. H., Patel, J., Kapadia, M., & Lakdawala, N. (2024) · Neurocomputing · doi:10.1016/j.neucom.2024.128284 · Cited by 48
End-to-End Video Text Spotting with Transformer
Wu, W., Cai, Y., Shen, C., Zhang, D., Fu, Y., Zhou, H., & Luo, P. (2024) · International Journal of Computer Vision · doi:10.1007/s11263-024-02063-1 · Cited by 44
Self-citationYang, C., Chen, M., Yuan, Y., & Wang, Q. (2023) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2023.3244322 · Cited by 38
“Different from the above works, Wang et al. [24], [25], [26] and Liao et al. [27], [28] proposed expansion strategies to generate text regions from shrink regions, which avoided detecting multiple adhesive texts as one either.”
ResultRecurrent Generic Contour-Based Instance Segmentation With Progressive Learning
Feng, H., Zhou, K., Zhou, W., Yin, Y., Deng, J., Sun, Q., & Li, H. (2024) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2024.3383829 · Cited by 39
SwinTextSpotter v2: Towards Better Synergy for Scene Text Spotting
Huang, M., Peng, D., Li, H., Peng, Z., Liu, C., Lin, D., Liu, Y., Bai, X., & Jin, L. (2025) · International Journal of Computer Vision · doi:10.1007/s11263-025-02428-0 · Cited by 31
“An additional advantage of an end-to-end system is that it is simpler to maintain and transfer to new domains, compared to a cascaded pipeline, which is tightly connected to the data and requires significant engineering efforts (Liu et al., 2021a; Wang et al., 2021d).”
Turning a CLIP Model Into a Scene Text Spotter
Yu, W., Liu, Y., Zhu, X., Cao, H., Sun, X., & Bai, X. (2024) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2024.3379828 · Cited by 34
Natural scene text detection and recognition based on saturation-incorporated multi-channel MSER
Tong, G., Dong, M., Sun, X., & Song, Y. (2022) · Knowledge-Based Systems · doi:10.1016/j.knosys.2022.109040 · Cited by 26
Boundary TextSpotter: Toward Arbitrary-Shaped Scene Text Spotting
Lu, P., Wang, H., Zhu, S., Wang, J., Bai, X., & Liu, W. (2022) · IEEE Transactions on Image Processing · doi:10.1109/tip.2022.3206615 · Cited by 27
Hierarchical Multiview Top-k Pooling With Deep-Q-Networks
Li, Z., Su, H., Wu, Y., Zhang, Q., Yuan, C., Gribova, V. V., Filaretov, V. F., & Huang, D.-S. (2023) · IEEE Transactions on Artificial Intelligence · doi:10.1109/tai.2023.3334261 · Cited by 24
Decoupling Recognition from Detection: Single Shot Self-Reliant Scene Text Spotter
Wu, J., Lyu, P., Lu, G., Zhang, C., Yao, K., & Pei, W. (2022) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3503161.3548266 · Cited by 30
“Following the routine training paradigm adopted by most exiting text spotters [22, 26, 36], we first pre-train our SRSTS on two synthetic datasets (‘Synthtext’ [10] and ‘Bezier Curve’ [21]), then finetune the model on a mixture of synthetic and real-world datasets.”
MethodInverse-Like Antagonistic Scene Text Spotting via Reading-Order Estimation and Dynamic Sampling
Zhang, S.-X., Yang, C. H., Zhu, X., Zhou, H., Wang, H., & Yin, X.-C. (2024) · IEEE Transactions on Image Processing · doi:10.1109/tip.2024.3352399 · Cited by 22
“…Zhu, Hongyang Zhou, and Xu-Cheng Yin are with the School of Computer and Communication Engineering, University of Science and Technology Beijing (USTB), Beijing 100083, China (e-mail: zhangshixue111@163.com; chunyang@ustb.edu.cn; zhuxiaobin@ ustb.edu.cn; ilaopi@foxmail.com; xuchengyin@ustb.edu.cn).”
BackgroundInfluentialA decade: Review of scene text detection methods
Rainarli, E., Suprapto, S., & Wahyono, W. (2021) · Computer Science Review · doi:10.1016/j.cosrev.2021.100434 · Cited by 23
CBNet: A Plug-and-Play Network for Segmentation-Based Scene Text Detection
Zhao, X. Q., Feng, W., Zhang, Z., Lv, J., Zhu, X., Lin, Z., Hu, J., & Shao, J. (2024) · International Journal of Computer Vision · doi:10.1007/s11263-024-02022-w · Cited by 26
DNTextSpotter: Arbitrary-Shaped Scene Text Spotting via Improved Denoising Training
Qiao, Q., Xie, Y., Gao, J., Wu, T., Huang, S., Fan, J., Cao, Z., Wang, Z., & Zhang, Y. (2024) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3664647.3680981 · Cited by 32
“Text Spotter, as an essential foundational technology that encompasses text detection and recognition, plays a critical role in various domains [6, 32, 39, 49] such as autonomous driving, security monitoring, and social media analysis.”
A DCNN-based arbitrarily-oriented object detector with application to quality control and inspection
Yao, K., Ortiz, A., & Bonnin‐Pascual, F. (2022) · Computers in Industry · doi:10.1016/j.compind.2022.103737 · Cited by 18
“With the advent of DCNN-based methodologies, a number of objecte detection approaches have emerged in a number of domains, including remote sensing [17, 18, 19], text recognition [20, 21, 22, 23], face detection [24, 25, 26], or surveillance [27, 28, 29], to name but a few.”
FastTextSpotter: A High-Efficiency Transformer for Multilingual Scene Text Spotting
Das, A., Biswas, S., Pal, U., Lladós, J., & Bhattacharya, S. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-78498-9_10 · Cited by 17
Focus Entirety and Perceive Environment for Arbitrary-Shaped Text Detection
Han, X., Gao, J., Yang, C., Yuan, Y., & Wang, Q. (2024) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2024.3521797 · Cited by 14
“PAN++ [50] surpasses ours 0.2 % in F-measure, which is mainly because it uses ICDAR2017-MLT to pre-train, but the proposed method uses SynthText.”
MethodInfluentialSpotlight Text Detector: Spotlight on Candidate Regions Like a Camera
Han, X., Gao, J., Yang, C., Yuan, Y., & Wang, Q. (2024) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2024.3521824 · Cited by 14
“Specifically, P, R, and F can be represented as follows:”
Weakly supervised scene text generation for low-resource languages
Xie, Y., Chen, X., Zhan, H., Shivakumara, P., Yin, B., Liu, C., & Lu, Y. (2023) · Expert Systems with Applications · doi:10.1016/j.eswa.2023.121622 · Cited by 17
Diving into the Depths of Spotting Text in Multi-Domain Noisy Scenes
Das, A., Biswas, S., Pal, U., & Lladós, J. (2024) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra57147.2024.10611120 · Cited by 15
“However, they were computationally more expensive and PAN++ [52] improved its efficiency by using a faster detector to multiply the segmentation maps with the features to suppress the background as followed in other approaches [53], [40].”
MethodDiscriminative feature selection with directional outliers correcting for data classification
Yuan, L., Yang, G., Xu, Q., & Lü, T. (2022) · Pattern Recognition · doi:10.1016/j.patcog.2022.108541 · Cited by 13
Lian, Z., Yin, Y., Zhi, M., & Xu, Q. (2023) · Electronics · doi:10.3390/electronics12143055 · Cited by 14
Das, A., Biswas, S., Banerjee, A., Lladós, J., Pal, U., & Bhattacharya, S. (2024) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv57701.2024.00077 · Cited by 14
Xu, J., Lin, A., Li, J., & Lu, G. (2023) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2023.3285096 · Cited by 11
Real-Time Text Detection With Similar Mask in Traffic, Industrial, and Natural Scenes
Han, X., Gao, J., Yang, C., Yuan, Y., & Wang, Q. (2024) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2024.3485061 · Cited by 11
An Effective Method for Detection and Recognition of Uyghur Texts in Images with Backgrounds
Ibrayim, M., Mattohti, A., & Hamdulla, A. (2022) · Information · doi:10.3390/info13070332 · Cited by 11
Bridging the Gap Between End-to-End and Two-Step Text Spotting
Huang, M., Li, H., Liu, Y., Bai, X., & Jin, L. (2024) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52733.2024.01478 · Cited by 22
“In comparison, [44, 56] use the RoI Masking to connect the detector and recognizer.”
MethodTexts as points: Scene text detection with point supervision
Zhao, M., Feng, W., Yin, F., & Liu, C. (2023) · Pattern Recognition Letters · doi:10.1016/j.patrec.2023.04.004 · Cited by 11
TextBlockV2: Towards Precise-Detection-Free Scene Text Spotting with Pre-trained Language Model
Lyu, J., Jin, W., Zeng, G., Zeng, L., Xie, E., Wang, W., Ma, C., & Zhou, Y. (2025) · ACM Transactions on Multimedia Computing Communications and Applications · doi:10.1145/3734872 · Cited by 10
“Moreover, a few methods [35]–[37] adopt bottom-up paradigms for scene text detection.”
Self-citationShu, Y., Wang, W., Zhou, Y., Liu, S., Zhang, A., Yang, D., & Wang, W. (2023) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3581783.3612383 · Cited by 10
“To tackle the challenge of curved text detection, segmentation-based methods [8, 60, 78, 84] have been proposed.”
MethodSelf-citationComprehensive benchmark datasets for Amharic scene text detection and recognition
Dikubab, W., Liang, D., Liao, M., & Bai, X. (2022) · Science China Information Sciences · doi:10.1007/s11432-021-3447-9 · Cited by 10
Self-citationGranularity-Aware Single-Point Scene Text Spotting With Sequential Recurrence Self-Attention
Tong, X., Dai, P., Qin, X., Wang, R., & Ren, W. (2024) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2024.3431993 · Cited by 9
“For horizontal, multi-oriented, and arbitrarily-shaped scene texts, scholars have designed different ROI alignment mechanisms, e.g., ROI Bilinear Sampling [4], ROIRotate [5], BezierAlign [6], ROI Masking [7], etc.”
TextFormer: A Query-based End-to-end Text Spotter with Mixed Supervision
Zhai, Y., Zhang, X., Qin, X., Zhao, S., Dong, X., & Shen, J. (2024) · Machine Intelligence Research · doi:10.1007/s11633-023-1460-6 · Cited by 9
DMCTDet: A density map-guided composite transformer network for object detection of UAV images
Li, J., Guo, S., Yi, S., He, R., & Jia, Y. (2025) · Signal Processing Image Communication · doi:10.1016/j.image.2025.117284 · Cited by 10
Kernel-mask knowledge distillation for efficient and accurate arbitrary-shaped text detection
Chen, H., Qiu, Y., Jiang, M., Lin, J., & Chen, P. (2023) · Complex & Intelligent Systems · doi:10.1007/s40747-023-01134-z · Cited by 9
MIEP: Channel Pruning with Multi-granular Importance Estimation for Object Detection
Jiang, L., Chen, J., Huang, D., & Wang, Y. (2023) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3581783.3612563 · Cited by 11
“Object detection [12, 37, 50, 65, 84] is a fundamental task in computer vision and has a wide range of applications in multimedia content understanding such as autonomous driving [3, 46], cross-modal retrieval [53, 61], image caption [1, 74], and optical character recognition [60, 63].”
BackgroundGoMatching++: Parameter- and Data-Efficient Arbitrary-Shaped Video Text Spotting and Benchmarking
He, H., Zhang, J., Ye, M., Liu, J., Du, B., & Tao, D. (2026) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2026.3688478 · Cited by 7
“Early approaches [30], [31], [32] crafted RoI-based modules to bridge text detection and recognition.”
Wu, W., Li, Z., Cai, Y., Zhou, H., & Shou, M. Z. (2024) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2024.3454331 · Cited by 11
“The architecture includes backbone, FPN, up-sample, Masked roi [33] and three heads, i.e., detection, recognition, and track heads.”
InfluentialAn End-to-End Scene Text Recognition for Bilingual Text
Albalawi, B. M., Jamal, A. T., Khuzayem, L. A., & Alsaedi, O. A. (2024) · Big Data and Cognitive Computing · doi:10.3390/bdcc8090117 · Cited by 8
A survey on methods, datasets and implementations for scene text spotting
Blanco‐Medina, P., Fidalgo, E., Alegre, E., & González‐Castro, V. (2022) · IET Image Processing · doi:10.1049/ipr2.12574 · Cited by 8
An accurate approach to real-time machine-readable zone detection with mobile devices
Gayer, A. V., Ershova, D. M., & Arlazarov, V. V. (2023) · International Journal on Document Analysis and Recognition (IJDAR) · doi:10.1007/s10032-023-00435-w · Cited by 7
“Another models like TextSpotter [19], Mask TextSpotter [20], CharNet [21] and PAN++ [22], allow not only to detect text, but also to recognize it at the same time.”
MethodDynamic Low-Resolution Distillation for Cost-Efficient End-to-End Text Spotting
Chen, Y., Liang, Q., Cheng, Z., Pu, S., Niu, Y., & Li, X. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-19815-1_21 · Cited by 10
Self-citationDetection and Identification of Text-based Traffic Signs
Chi, X., Luo, D., Liang, Q., Yang, J., & Huang, H. (2023) · Sensors and Materials · doi:10.18494/sam4253 · Cited by 6
Universal Defensive Underpainting Patch: Making Your Text Invisible to Optical Character Recognition
Deng, J., Dong, L., Chen, J., Yan, D., Wang, R., Ye, D., Zhao, L., & Tian, J. (2023) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3581783.3613768 · Cited by 6
“We take state-of-the-art semantic segmentation models CRAFT, DBnet, PAN++[18], PSENet[17] and EasyOCR as exem-plar attacking targets to develop UDUP, which aids in protecting copyrighted texts by helping them escape detection from STDs.”
MethodVimTS: A Unified Video and Image Text Spotter for Enhancing the Cross-Domain Generalization
Liu, Y., Huang, M., Yan, H., Deng, L., Wu, W., Lü, H., Shen, C., Jin, L., & Bai, X. (2025) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2025.3528950 · Cited by 6
“The output of the REM and ResNet are sent into a Transformer encoder [27] to further enhance the long-range dependencies across various scales.”
Self-citationCommuSpotter: Scene Text Spotting with Multi-Task Communication
Zhao, L., Wilsbacher, G., & Wang, S. (2023) · Applied Sciences · doi:10.3390/app132312540 · Cited by 6
EK-Net:Real-Time Scene Text Detection with Expand Kernel Distance
Zhu, B., Liu, F., Chen, X., & Tang, Q. (2024) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp48485.2024.10448354 · Cited by 6
“PAN [19] and PAN++ [20] introduce a low computational-cost segmentation head and a learnable post-processing technique based on PSENet [7].”
MethodVGTS: Visually Guided Text Spotting for novel categories in historical manuscripts
Hu, W., Zhan, H., Ma, X., Liu, C., Yin, B., Lu, Y., & Suen, C. Y. (2024) · Expert Systems with Applications · doi:10.1016/j.eswa.2024.125557 · Cited by 8
TextMamba: Scene Text Detector with Mamba
Zhao, Q., Yan, Y. P., & Wang, D.-H. (2025) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn64981.2025.11227846 · Cited by 9
“For example, PSE [1] and PAN [3], [5] predict text kernels and apply adaptive scaling algorithms.”
ACP-Net: Asymmetric Center Positioning Network for Real-Time Text Detection
Zhu, B., Liu, F., Chen, X. T., Tang, Q., & Chen, C. (2024) · Knowledge-Based Systems · doi:10.1016/j.knosys.2024.112603 · Cited by 5
Wavelet descriptor network for arbitrary-shaped text detection
Zhang, Z., & Tong, M. (2022) · Journal of Electronic Imaging · doi:10.1117/1.jei.31.4.043051 · Cited by 5
When IC meets text: Towards a rich annotated integrated circuit text dataset
Ng, C. C., Lin, C.-T., Tan, Z. Q., Wang, X., Kew, J. L., Chan, C. S., & Zach, C. (2023) · Pattern Recognition · doi:10.1016/j.patcog.2023.110124 · Cited by 5
Huang, L., Liao, S., & Yang, W. (2023) · The Visual Computer · doi:10.1007/s00371-023-03093-5 · Cited by 5
Road Scene Text Detection and Recognition Using Machine Learning
Shah, S. H. A., & Shah, J. H. (2023) · IEEE International Conference on Smart Communities: Improving Quality of Life using AI, Robotics and IoT (HONET) · doi:10.1109/honet59747.2023.10375032 · Cited by 7
“Besides these older approaches, Deep neural networks have also contributed significantly in some cases [17-20].”
BackgroundBIPOOLNET: An advanced UNet architecture for enhanced lane detection in autonomous vehicles
Santhiya, P., Jebadurai, I. J., Paulraj, G. J. L., & Jenefa, A. (2024) · Intelligent Decision Technologies · doi:10.3233/idt-240162 · Cited by 4
“The FPEM module, featuring a cascadable U-shape architecture, enhances the encoder module by adding multi-level information to aid in improved segmentation [21,22].”
A real-time arbitrary-shape text detector
Lu, M., Li, L., & Chen, C. (2024) · PLoS ONE · doi:10.1371/journal.pone.0302234 · Cited by 4
LTDNet: A Lightweight Text Detector for Real-Time Arbitrary-Shape Traffic Text Detection
Wang, R., Zhu, Y., Zhu, Z., Cui, L., Wan, Z., Zhu, A., Ding, Y., Qian, S., Gao, C., & Sang, N. (2025) · IEEE/CAA Journal of Automatica Sinica · doi:10.1109/jas.2024.125022 · Cited by 4
“Our method compared with the PAN++ [17] (Baseline) on three traffic text datasets CTST-1600 [31], ODETT-3000 and TPD [44], as well as two scene text datasets CTW1500 [45] and IC15 [45]. been conducted on three traffic text datasets and two scene text datasets to demonstrate the superiority ofLTDNet.”
InfluentialTTS: Hilbert Transform-Based Generative Adversarial Network for Tattoo and Scene Text Spotting
Banerjee, A., Shivakumara, P., Pal, U., Antonacopoulos, A., Lü, T., & Lladós, J. (2024) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2024.3378458 · Cited by 5
DiZNet: An end-to-end text detection and recognition algorithm with detail in text zone
Zhou, D., Zhang, J., & Li, C. (2024) · Journal of Visual Communication and Image Representation · doi:10.1016/j.jvcir.2024.104261 · Cited by 4
Artistic-style text detector and a new Movie-Poster dataset
Ning, A., Xue, M., Wei, Y., Zhou, M., & Zhong, S. (2024) · Expert Systems with Applications · doi:10.1016/j.eswa.2024.125544 · Cited by 4
Text Font Correction and Alignment Method for Scene Text Recognition
Ding, L., Liu, Y., Zhao, Q., & Liu, Y. (2024) · Sensors · doi:10.3390/s24247917 · Cited by 5
Few-shot Shape Recognition by Learning Deep Shape-aware Features
Shi, W., Lu, C., Shao, M., Zhang, Y., Xia, S., & Koniusz, P. (2024) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv57701.2024.00186 · Cited by 5
“In arbitrary-shaped text detection [13–16], the aim is to capture regions around various text shapes rather than using bounding boxes.”
BackgroundFew-shot cross-modal text detection via CLIP
Zhao, Q., Liang, L., Zhang, X., Zhang, T., Li, J., & Liu, Y. (2025) · The Visual Computer · doi:10.1007/s00371-025-03962-1 · Cited by 3
Natural Scene Text Detection Algorithm via Internal Feature Enhancement and Adaptive Cross Fusion
Wang, G., Wei, S., Yang, D., & Guo, A. (2025) · IEEE Access · doi:10.1109/access.2025.3602805 · Cited by 3
Pull Pole Points to Text Contour by Magnetism: A Real-Time Scene Text Detector
Han, X., Yang, C., & Wang, Q. (2025) · IEEE Transactions on Image Processing · doi:10.1109/tip.2025.3609196 · Cited by 3
“1) Shrink Mask: PAN [36], and DBNet ++ [30] utilize the Vatti clipping algorithm [37] to generate shrink masks.”
InfluentialVertically invariant network for vertical and horizontal scene text recognition
Zhu, C., Feng, Z., He, J., & Xiao, X. (2025) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2025.112680 · Cited by 3
Masked Text Pre-Training for Scene Text Detection
Xie, H., Wang, K., Zhou, B., Wang, Y., Qi, W., Wu, J., Qu, Y., Gao, Z., Zhang, D., & Liu, Y. (2025) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2025.3613181 · Cited by 4
LRANet++: Low-Rank Approximation Network for Accurate and Efficient Text Spotting
Su, Y., Chen, Z., Du, Y., Wu, Z., Xie, H., & Jiang, Y. (2026) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2026.3650769 · Cited by 3
“Furthermore, ROI-Mask [37] is naturally suited to the mask-driven method to extract features for recognition.”
AAF-Net: Scene text detection based on attention aggregation features
Chen, M., Ibrayim, M., & Hamdulla, A. (2022) · PLoS ONE · doi:10.1371/journal.pone.0272322 · Cited by 3
“PAN++ [8] is an end-to-end scene text detection and recognition network.”
BackgroundDPGS: Cross-cooperation guided dynamic points generation for scene text spotting
Sun, W., Wang, Q., Hou, Z., Chen, X., Yan, Q., & Zhang, Y. (2024) · Knowledge-Based Systems · doi:10.1016/j.knosys.2024.112399 · Cited by 3
Double supervision for scene text detection and recognition based on BMINet
Wan, H., Liu, R., & Li, Y. (2024) · Signal Processing Image Communication · doi:10.1016/j.image.2024.117226 · Cited by 3
Overview of Scene Text Detection and Recognition
Kang, J., Ibrayim, M., & Hamdulla, A. (2022) · International Conference on Measuring Technology and Mechatronics Automation (ICMTMA) · doi:10.1109/icmtma54903.2022.00137 · Cited by 3
EWST: an extreme weather scene text detector with dehazing and localization refinement
Liu, Y., Wang, R., Zhu, G., Liu, M., Han, C., He, X., & Gao, C. (2023) · Journal of Electronic Imaging · doi:10.1117/1.jei.32.1.013007 · Cited by 3
FTDNet: Joint Semantic Learning for Scene Text Detection in Adverse Weather Conditions
Tian, J., Zhou, G., Liu, Y., Deng, E., & Jia, Z. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-41734-4_9 · Cited by 3
Advanced License Plate Detector in Low-Quality Images with Smooth Regression Constraint
Yu, J., Liu, D., Wang, T., Tian, J., Xu, F., & Cao, J. (2023) · Lecture notes in computer science · doi:10.1007/978-981-99-8432-9_30 · Cited by 3
EK-Net++: Real-time scene text detection with expand kernel distance and Epoch Adaptive Weight
Zhu, B., Chen, X. T., Tang, Q., Chen, C., & Liu, F. (2024) · Expert Systems with Applications · doi:10.1016/j.eswa.2024.126159 · Cited by 6
A Lightweight Context-Driven Training-Free Network for Scene Text Segmentation and Recognition
Chakraborty, R., Shivakumara, P., Pal, U., & Liu, C. (2025) · Lecture notes in computer science · doi:10.1007/978-3-032-04617-8_15 · Cited by 3
“Recent models prioritized towards efficiency such as [48,60,61].”
A novel domain independent scene text localizer
Roy, A., Shivakumara, P., Pal, U., & Liu, C. (2024) · Pattern Recognition · doi:10.1016/j.patcog.2024.111015 · Cited by 2
Text-Pass Filter: An Efficient Scene Text Detector
Yang, C., Ma, H., Han, X., Yuan, Y., & Wang, Q. (2026) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2026.3668365 · Cited by 2
“Considering the low efficiency of the stepwise expansion process, Wang et al. [46], [47] predicted shrink-masks with one specific scale, and whole regions then were clustered into shrink-masks according to the idea of the clustering algorithm.”
Scene Text Detection Based on Multi-Headed Self-Attention Using Shifted Windows
Huang, B., & Feng, X. (2023) · Applied Sciences · doi:10.3390/app13063928 · Cited by 2
ADATS: Adaptive RoI-Align based Transformer for End-to-End Text Spotting
Huang, Z., Wan, Q., Chen, J., Zhao, X., Ye, K., & Shen, L. (2023) · IEEE International Conference on Multimedia and Expo (ICME) · doi:10.1109/icme55011.2023.00243 · Cited by 2
Chinese Text Spotter Exploiting Spatial Semantic Information in Scene Text Images
Wang, H., & Zhou, H. (2023) · International Conference on Robotics and Computer Vision (ICRCV) · doi:10.1109/icrcv59470.2023.10329042 · Cited by 2
“On the other hand, research by [21]–[24] proposed character-level arbitrary-shaped text detection and recognition methods without the need for RoI operations.”
BackgroundSynthetic Data Generation for Text Spotting on Printed Circuit Board Component Images
Liau, W. J. B., Tay, S. C., Mohamed, A. S. A., Wahab, M. N. A., Lim, L. C., Kang, K. B., & Noor, M. H. M. (2024) · IEEE Access · doi:10.1109/access.2024.3394524 · Cited by 2
“Generally, two-staged text spotters sequentially detect and recognize text in an image [2], [3], [4], [5], [41], [42], [43], [44], [45], [46], [47], [48], [49], [50], [51], [52], [53], [54].”
BackgroundMethodInfluentialImproving Scene Text Retrieval via Stylized Middle Modality
Zhu, S., Fang, J. P., Fang, P., & Xue, H. (2024) · ACM Transactions on Multimedia Computing Communications and Applications · doi:10.1145/3696209 · Cited by 2
“PAN++ [41] leverages the lightweight module to detect text kernel instead of complex segmentation.”
Xie, Y., Xu, C., Shi, C., Li, J., Yuan, Z., & Qiao, Q. (2024) · Lecture notes in computer science · doi:10.1007/978-981-96-0119-6_42 · Cited by 4
Wang, H., Zhou, H., Zhang, Y., Ma, J., & Ling, H. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3560748 · Cited by 3
“Current algorithms [8], [9], [10], [11], [12], [13] often overlook this imbalance and tend to prioritize learning from “head samples”—words of regular length that frequently appear—while neglecting “tail samples” of rare, extreme lengths, resulting in suboptimal training for the extremes of the…”
Influential
Metadata
- Authors
- Wang, W., Xie, E., Li, X., Liu, X., Liang, D., Zhibo, Y., Lu, T., & Shen, C.
- Affiliations
- Nanjing University (CN)
- University of Hong Kong (HK)
- Nanjing University of Science and Technology (CN)
- Sensetime (China) (CN)
- Monash University (AU)
- Venue
- IEEE Transactions on Pattern Analysis and Machine Intelligence
- Published
- 2021
- Volume
- 44
- Issue
- 9
- Pages
- 1-1
- ISSN
- 0162-8828 (ISSN-L)
- 2160-9292
- 1939-3539
- OpenAlex
- W3159307593
- Semantic Scholar
- 70aa3857040d2934ab7544988468ede697cb90c9
- Cited by
- 140
- References
- 73
- Citations by year
- 2026: 6 · 2025: 31 · 2024: 47 · 2023: 27 · 2022: 26 · 2021: 3
- FWCI
- 9.26
- Type
- journal article
- Language
- en
- Open access
- closed
- Locations
- IEEE Transactions on Pattern Analysis and Machine Intelligence (publishedVersion)
- PubMed (publishedVersion)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Handwritten Text Recognition Techniques
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Vehicle License Plate Recognition
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Keywords
- scene text detection, scene text recognition, lightweight detection head, pixel aggregation, real-time inference
- MeSH terms
- Algorithms
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.