Show, Attend and Read: A Simple and Strong Baseline for Irregular Text Recognition

Recognizing irregular text in natural scene images is challenging due to the large variance in text appearance, such as curvature, orientation and distortion. Most existing approaches rely heavily on sophisticated model designs and/or extra fine-grained annotations, which, to some extent, increase the difficulty in algorithm implementation and data collection. In this work, we propose an easy-to-implement strong baseline for irregular scene text recognition, using offthe-shelf neural network components and only word-level annotations. It is composed of a 31-layer ResNet, an LSTMbased encoder-decoder framework and a 2-dimensional attention module. Despite its simplicity, the proposed method is robust. It achieves state-of-the-art performance on irregular text recognition benchmarks and comparable results on regular text datasets. The code will be released.

• Publication year

  2018

• Venue

  AAAI Conference on Artificial Intelligence

• Publication date

  2018-11-02

• Fields of study

  Computer Science

• Identifiers
  DOI 10.1609/aaai.v33i01.33018610arXiv 1811.00751
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• ASTER: An Attentional Scene Text Recognizer with Flexible Rectification

  2019influential reference
• Edit Probability for Scene Text Recognition

  2018influential reference
• Char-Net: A Character-Aware Neural Network for Distorted Scene Text Recognition

  2018cited by this paper
• SqueezedText: A Real-Time Scene Text Recognition by Binary Convolutional Encoder-Decoder Network

  2018cited by this paper
• AON: Towards Arbitrarily-Oriented Text Recognition

  2017influential reference
• Learning to Read Irregular Text with Attention Mechanisms

  2017influential reference
• Gated Recurrent Convolution Neural Network for OCR

  2017influential reference
• Towards End-to-End Text Spotting with Convolutional Recurrent Neural Networks

  2017cited by this paper
• Focusing Attention: Towards Accurate Text Recognition in Natural Images

  2017cited by this paper
• Recursive Recurrent Nets with Attention Modeling for OCR in the Wild

  2016cited by this paper
• COCO-Text: Dataset and Benchmark for Text Detection and Recognition in Natural Images

  2016cited by this paper
• STAR-Net: A SpaTial Attention Residue Network for Scene Text Recognition

  2016influential reference
• Robust Scene Text Recognition with Automatic Rectification

  2016influential reference
• Synthetic Data for Text Localisation in Natural Images

  2016cited by this paper
• An End-to-End Trainable Neural Network for Image-Based Sequence Recognition and Its Application to Scene Text Recognition

  2015cited by this paper
• Show, Attend and Tell: Neural Image Caption Generation with Visual Attention

  2015influential reference
• Attention-Based Models for Speech Recognition

  2015cited by this paper
• Spatial Transformer Networks

  2015cited by this paper
• Reading Scene Text in Deep Convolutional Sequences

  2015influential reference
• Deep Residual Learning for Image Recognition

  2015influential reference
• Sequence to Sequence Learning with Neural Networks

  2014cited by this paper
• Strokelets: A Learned Multi-scale Representation for Scene Text Recognition

  2014cited by this paper
• Adam: A Method for Stochastic Optimization

  2014cited by this paper
• A robust arbitrary text detection system for natural scene images

  2014cited by this paper
• Reading Text in the Wild with Convolutional Neural Networks

  2014influential reference
• ICDAR 2013 Robust Reading Competition

  2013influential reference
• Recognizing Text with Perspective Distortion in Natural Scenes

  2013influential reference
• Top-down and bottom-up cues for scene text recognition

  2012cited by this paper
• End-to-end scene text recognition

  2011influential reference
• Scene Text Recognition using Higher Order Language Priors

  2009cited by this paper
• NewsWeeder: Learning to Filter Netnews

  1995cited by this paper

• Summarize Before Glimpse: Brain-Inspired Non-Autoregressive Scene Text Recognizer

  2026cites this paper
• End-to-End Shipping Container Code Spotting Based on Visual Sensors for Intelligent Inspection

  2026cites this paper
• ETLRS: An End-to-End Text Spotting Framework With Spatial and Directional Perception and Cross-Task Feature Synergy

  2026cites this paper
• Improving Chinese Text Recognition with Multi-Granularity Features and Vision-Language Reasoning

  2026cites this paper
• What Is Wrong with Synthetic Data for Scene Text Recognition? A Strong Synthetic Engine with Diverse Simulations and Self-Evolution

  2026cites this paper
• Advanced Methods for Text Detection and Recognition in Customs Declaration Forms: The EFF-DBNet and SKSVTR

  2025cites this paper
• Reimagining CRNN with Attention for Handwritten Chinese Text Recognition in Noisy Backgrounds

  2025cites this paper
• Attention-based transformer models for image captioning across languages: An in-depth survey and evaluation

  2025cites this paper
• Lightweight Swin-Transformer and DistilGPT-2 for Image-to-LaTeX in Constrained Environments

  2025cites this paper
• PLATTER: A Page-Level Handwritten Text Recognition System for Indic Scripts

  2025cites this paper
• From Broadcast to Minimap: Achieving State-of-the-Art Soccernet Game State Reconstruction

  2025cites this paper
• Efficient Scene Text Recognition in Noisy Environments Using Fusion-Based Adaptation and Triple-Level Confidence Modeling

  2025cites this paper
• Skeleton-Guided Artistic Text Recognition

  2025cites this paper
• Font-Agent: Enhancing Font Understanding with Large Language Models

  2025cites this paper
• LVM-OCR: A Transformer-Based Architecture for Context-Aware Document Understanding

  2025cites this paper
• UniRec-0.1B: Unified Text and Formula Recognition with 0.1B Parameters

  2025cites this paper
• UniMTR: Unified Recognition of Dual-style Traditional Mongolian Scripts via Contrastive Representation Alignment

  2025cites this paper
• MDiff4STR: Mask Diffusion Model for Scene Text Recognition

  2025cites this paper
• SDRT: Enhance Vision-Language Models by Self-Distillation with Diverse Reasoning Traces

  2025cites this paper
• IMTLM-Net: improved multi-task transformer based on localization mechanism network for handwritten English text recognition

  2025cites this paper
• Chinese Character Recognition based on Swin Transformer-Encoder

  2025cites this paper
• Water Meter Reading Based on Text Recognition Techniques and Deep Learning

  2025cites this paper
• An attention-guided image super-resolution for scene text

  2025cites this paper
• CCDPlus: Towards Accurate Character to Character Distillation for Text Recognition

  2025cites this paper
• A Lightweight Multi-Module Fusion Approach for Korean Character Recognition

  2025cites this paper
• AI for Empowering Disabilities

  2025cites this paper
• MetaWriter: Personalized Handwritten Text Recognition Using Meta-Learned Prompt Tuning

  2025cites this paper
• Enhanced Chinese scene text recognition model base on cross-domain feature fusion

  2025cites this paper
• Multi‐Task Learning for Chinese Character and Radical Recognition With Dynamic Channel‐Spatial Attention and Rotational Positional Encoding

  2025cites this paper
• From Fragment to One Piece: A Review on AI-Driven Graphic Design

  2025cites this paper
• Robust Automatic License Plate Recognition Using Synthetic Data and Transformer-Based Deep Learning Model

  2025cites this paper
• Automatic License Plate Recognition in In-the-Wild Scenarios: A Comprehensive Review, Open Issues, and Future Directions

  2025cites this paper
• Text-Aware Image Restoration with Diffusion Models

  2025cites this paper
• A Comprehensive Survey of Transformers in Text Recognition: Techniques, Challenges, and Future Directions

  2025cites this paper
• Vertically invariant network for vertical and horizontal scene text recognition

  2025cites this paper
• Zero-shot Artistic Text Recognition with Multimodal Language Models

  2025cites this paper
• Integrating vision and language: a novel approach to translation for low-resource Indic languages

  2025cites this paper
• LICS: Locating Inter-Character Spaces for Multilingual Scene Text Detection

  2025influential citation
• FastGOT: Task-Driven Visual Tokens Pruning

  2025cites this paper
• SmartExp: An Adaptive Data Expansion Strategy for Improving Handwritten Text Recognition

  2025cites this paper
• Enforcing Graph Structures to Enhance Key Information Extraction in Document Analysis

  2025cites this paper
• A robust pointer meter reading method for inspection robots in real industrial scenarios

  2025cites this paper
• Attention Guidance by Cross-Domain Supervision Signals for Scene Text Recognition

  2025cites this paper
• SDRNet: A hybrid approach with deep convolutional networks for ship draft reading

  2025cites this paper
• From Chaotic OCR Words to Coherent Document: A Fine-to-Coarse Zoom-Out Network for Complex-Layout Document Image Translation

  2025cites this paper
• A deep learning-based pipeline for the conservation assessment of bindings in archives and libraries

  2025cites this paper
• From Detection to Understanding: A Systematic Survey of Deep Learning for Scene Text Processing

  2025cites this paper
• Improving Irregular Text Recognition with Adaptive Feature Compression

  2025cites this paper
• Adaptive Feature Aggregation for In-Air Handwritten Trajectory

  2025cites this paper
• Fast and Lightweight Automatic Shipping Container Attributes Spotting

  2025cites this paper
• From Fragment to One Piece: A Survey on AI-Driven Graphic Design

  2025cites this paper
• Optical Recognition of the Philippines' Ancient Text: A Deep Learning Approach

  2025cites this paper
• A novel scene text image super-resolution approach

  2025cites this paper
• Scene Chinese Recognition with Mix Attention and LLM

  2025cites this paper
• Feature enhanced attention decoder for scene text recognition

  2025cites this paper
• DOTA: Deformable Optimized Transformer Architecture for End-to-End Text Recognition with Retrieval-Augmented Generation

  2025cites this paper
• Transformer-Based Bidirectional Attention Network for Segmentation-Free Word-Level Text Recognition with Overlapping Characters

  2025cites this paper
• ClapperText: A Benchmark for Text Recognition in Low-Resource Archival Documents

  2025cites this paper
• Joint Low-level and High-level Textual Representation Learning with Multiple Masking Strategies

  2025cites this paper
• Exploring the effectiveness of data-centric AI approaches to developing a prescription recognition system

  2025cites this paper
• Development of OCR Service for Page-Level Recognition for Camera-Captured Document Images

  2025cites this paper
• On the Behavior of Contrastive Regularization in Improving Chinese Text Recognizer

  2025cites this paper
• MIRROR: Multi-scale iterative refinement for robust chinese text recognition

  2025cites this paper
• Data-Driven Container Marking Detection and Recognition System With an Open Large-Scale Scene Text Dataset

  2024cites this paper
• SARN: Script-Aware Recognition Network for scene multilingual text recognition

  2024cites this paper
• SoccerNet Game State Reconstruction: End-to-End Athlete Tracking and Identification on a Minimap

  2024cites this paper
• Image as a Language: Revisiting Scene Text Recognition via Balanced, Unified and Synchronized Vision-Language Reasoning Network

  2024cites this paper
• Toward Unified End-to-End License Plate Detection and Recognition for Variable Resolution Requirements

  2024influential citation
• Online Sequence-Based Deep Learning Approach for Metallic Debossed and Embossed Turbomachinery Blade Text Recognition Application

  2024cites this paper
• Irregular text block recognition via decoupling visual, linguistic, and positional information

  2024cites this paper
• TextBlockV2: Towards Precise-Detection-Free Scene Text Spotting with Pre-trained Language Model

  2024cites this paper
• Multimodal learning with only image data: A deep unsupervised model for street view image retrieval by fusing visual and scene text features of images

  2024cites this paper
• Multimodal Visual-Semantic Representations Learning for Scene Text Recognition

  2024cites this paper
• Cascade 2D attentional decoders with context-enhanced encoder for scene text recognition

  2024influential citation
• Scene text recognition with context-aware autonomous bidirectional iterative models

  2024cites this paper
• Instruction-Guided Scene Text Recognition

  2024cites this paper
• An advanced subway train localization system using a vision-based kilometer marker recognition-assisted multi-sensor fusion method

  2024cites this paper
• Compressed Vision Transformer for Scene Text Recognition

  2024cites this paper
• Efficiently Leveraging Linguistic Priors for Scene Text Spotting

  2024cites this paper
• Scene text visual question answering by using YOLO and STN

  2024cites this paper
• A Multi-angle Text Recognition Algorithm

  2024influential citation
• Inverse-Like Antagonistic Scene Text Spotting via Reading-Order Estimation and Dynamic Sampling

  2024cites this paper
• Text Font Correction and Alignment Method for Scene Text Recognition

  2024cites this paper
• Ancient Tamil inscription recognition using detect, recognize and labelling, interpreter framework of text method

  2024cites this paper
• Efficient Vehicle Tracking and Authentication with YOLOv8: Integrating Deep Learning and OCR for Real-Time License Plate Detection

  2024cites this paper
• TextSSR: Diffusion-based Data Synthesis for Scene Text Recognition

  2024cites this paper
• PuzText: Self-Supervised Learning of Permuted Texture Representation for Multilingual Text Recognition

  2024cites this paper
• Single visual model based on transformer for digital instrument reading recognition

  2024cites this paper
• Multi global context-aware transformer for ship name recognition in IoT

  2024cites this paper
• Policy-driven Auto-Augmentation with Distillment Rewards for Scene Text Recognition

  2024cites this paper
• MAGIC: Multi-granularity domain adaptation for text recognition

  2024cites this paper
• SVTRv2: CTC Beats Encoder-Decoder Models in Scene Text Recognition

  2024influential citation
• DLaVA: Document Language and Vision Assistant for Answer Localization with Enhanced Interpretability and Trustworthiness

  2024cites this paper
• Align, enhance and read: Scene Tibetan text recognition with cross-sequence reasoning

  2024cites this paper
• Efficient CRNN: Towards end-to-end low resource Urdu text recognition using depthwise separable convolutions and gated recurrent units

  2024cites this paper
• SRRNet: A Transformer Structure With Adaptive 2-D Spatial Attention Mechanism for Cell Phone-Captured Shopping Receipt Recognition

  2024cites this paper
• CFOR: Character-First Open-Set Text Recognition via Context-Free Learning

  2024cites this paper
• Trust Prophet or Not? Taking a Further Verification Step toward Accurate Scene Text Recognition

  2024cites this paper
• Using CRAFT algorithm-based OCR system to improve multilanguage text detection and recognition

  2024cites this paper
• Document Parsing Unveiled: Techniques, Challenges, and Prospects for Structured Information Extraction

  2024cites this paper