mlsanityData to 7 Oct 2026

On Recognizing Texts of Arbitrary Shapes with 2D Self-Attention

arXiv 1910.04396 ↗First posted 10 Oct 2019cs.CVcited by 59

Not on the map. Why?

Abstract (arXiv)

Scene text recognition (STR) is the task of recognizing character sequences in natural scenes. While there have been great advances in STR methods, current methods still fail to recognize texts in arbitrary shapes, such as heavily curved or rotated texts, which are abundant in daily life (e.g. restaurant signs, product labels, company logos, etc). This paper introduces a novel architecture to recognizing texts of arbitrary shapes, named Self-Attention Text Recognition Network (SATRN), which is inspired by the Transformer. SATRN utilizes the self-attention mechanism to describe two-dimensional (2D) spatial dependencies of characters in a scene text image. Exploiting the full-graph propagation of self-attention, SATRN can recognize texts with arbitrary arrangements and large inter-character spacing. As a result, SATRN outperforms existing STR models by a large margin of 5.7 pp on average in "irregular text" benchmarks. We provide empirical analyses that illustrate the inner mechanisms and the extent to which the model is applicable (e.g. rotated and multi-line text). We will open-source the code.

Key facts. Cited by 59 papers in the map (data through 7 Oct 2026). First posted on arXiv 10 Oct 2019.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 9; 2022: 21; 2023: 9; 2024: 14; 2025: 5; 2026: 1.

59 papers.

Sort:

By area:

  1. 2021TrOCR: Transformer-based Optical Character Recognition with Pre-trained Modelscited by 278
  2. 2022Scene Text Recognition with Permuted Autoregressive Sequence Modelscited by 148
  3. 2021Vision Transformer for Fast and Efficient Scene Text Recognitioncited by 81
  4. 2023Multi-modal In-Context Learning Makes an Ego-evolving Scene Text Recognizercited by 53
  5. 2021Rethinking Text Line Recognition Modelscited by 45
  6. 2022Reading and Writing: Discriminative and Generative Modeling for Self-Supervised Text Recognitioncited by 45
  7. 2022Toward Understanding WordArt: Corner-Guided Transformer for Scene Text Recognitioncited by 45
  8. 2023Revisiting Scene Text Recognition: A Data Perspectivecited by 45
  9. 2022Multi-Granularity Prediction for Scene Text Recognitioncited by 36
  10. 2021Multi-modal Text Recognition Networks: Interactive Enhancements between Visual and Semantic Featurescited by 34
  11. 2022Self-supervised Character-to-Character Distillation for Text Recognitioncited by 27
  12. 2022Self-supervised Implicit Glyph Attention for Text Recognitioncited by 25
38 more
  1. 2022Levenshtein OCRcited by 24
  2. 2023Context Perception Parallel Decoder for Scene Text Recognitioncited by 19
  3. 2021Transformer-Based Approach for Joint Handwriting and Named Entity Recognition in Historical documentscited by 17
  4. 2024Instruction-Guided Scene Text Recognitioncited by 17
  5. 2022Multimodal Semi-Supervised Learning for Text Recognitioncited by 16
  6. 2022Towards Open-Set Text Recognition via Label-to-Prototype Learningcited by 12
  7. 2022COO: Comic Onomatopoeia Dataset for Recognizing Arbitrary or Truncated Textscited by 12
  8. 2024Class-Aware Mask-Guided Feature Refinement for Scene Text Recognitioncited by 10
  9. 2022Pure Transformer with Integrated Experts for Scene Text Recognitioncited by 9
  10. 2023A Transformer-based Approach for Arabic Offline Handwritten Text Recognitioncited by 9
  11. 2023MSdocTr-Lite: A Lite Transformer for Full Page Multi-script Handwriting Recognitioncited by 7
  12. 2023Multi-Granularity Prediction with Learnable Fusion for Scene Text Recognitioncited by 7
  13. 2023An Empirical Study of Scaling Law for OCRcited by 7
  14. 2023Relational Contrastive Learning for Scene Text Recognitioncited by 6
  15. 2021Revisiting Classification Perspective on Scene Text Recognitioncited by 5
  16. 2021Data Augmentation for Scene Text Recognitioncited by 5
  17. 2024SVIPTR: Fast and Efficient Scene Text Recognition with Vision Permutable Extractorcited by 5
  18. 2024Zero-Shot Paragraph-level Handwriting Imitation with Latent Diffusion Modelscited by 5
  19. 2025Text-Aware Image Restoration with Diffusion Modelscited by 5
  20. 2022A Comprehensive Gold Standard and Benchmark for Comics Text Detection and Recognitioncited by 4
  21. 2024Self-Supervised Learning for Text Recognition: A Critical Surveycited by 3
  22. 2021Why You Should Try the Real Data for the Scene Text Recognitioncited by 2
  23. 2021TRIG: Transformer-Based Text Recognizer with Initial Embedding Guidancecited by 2
  24. 2022Comprehensive Benchmark Datasets for Amharic Scene Text Detection and Recognitioncited by 2
  25. 2022Unitail: Detecting, Reading, and Matching in Retail Scenecited by 2
  26. 2022Text Detection & Recognition in the Wild for Robot Localizationcited by 2
  27. 2022Vision-Language Adaptive Mutual Decoder for OOV-STRcited by 2
  28. 2024Watermark Text Pattern Spotting in Document Imagescited by 2
  29. 2024Sequential Visual and Semantic Consistency for Semi-supervised Text Recognitioncited by 2
  30. 2024The First Swahili Language Scene Text Detection and Recognition Datasetcited by 2
  31. 2022IterVM: Iterative Vision Modeling Module for Scene Text Recognitioncited by 1
  32. 2022Scene Text Recognition with Single-Point Decoding Networkcited by 1
  33. 2023Looking and Listening: Audio Guided Text Recognitioncited by 1
  34. 2024Dataset and Benchmark for Urdu Natural Scenes Text Detection, Recognition and Visual Question Answeringcited by 1
  35. 2025Why Stop at Words? Unveiling the Bigger Picture through Line-Level OCRcited by 1
  36. 2025GraDeT-HTR: A Resource-Efficient Bengali Handwritten Text Recognition System utilizing Grapheme-based Tokenizer and Decoder-only Transformercited by 1
  37. 2022Lane detection with Position Embeddingcited by 0
  38. 2022An Evaluation of OCR on Egocentric Datacited by 0

50 most cited of 59 All 59 (JSON)