Synthetic Data and Artificial Neural Networks for Natural Scene Text Recognition
arXiv:1406.2227
Abstract
In this work we present a framework for the recognition of natural scene text. Our framework does not require any human-labelled data, and performs word recognition on the whole image holistically, departing from the character based recognition systems of the past. The deep neural network models at the centre of this framework are trained solely on data produced by a synthetic text generation engine -- synthetic data that is highly realistic and sufficient to replace real data, giving us infinite amounts of training data. This excess of data exposes new possibilities for word recognition models, and here we consider three models, each one "reading" words in a different way: via 90k-way dictionary encoding, character sequence encoding, and bag-of-N-grams encoding. In the scenarios of language based and completely unconstrained text recognition we greatly improve upon state-of-the-art performance on standard datasets, using our fast, simple machinery and requiring zero data-acquisition costs.
References in corpus (2)
Cited by in corpus (190)
- AutoML: A Survey of the State-of-the-Art
- Multiple Object Recognition with Visual Attention
- Focusing Attention: Towards Accurate Text Recognition in Natural Images
- Object Detection in 20 Years: A Survey
- Smart Augmentation - Learning an Optimal Data Augmentation Strategy
- Rosetta: Large scale system for text detection and recognition in images
- Systematic evaluation of CNN advances on the ImageNet
- COCO-Text: Dataset and Benchmark for Text Detection and Recognition in Natural Images
- Use What You Have: Video Retrieval Using Representations From Collaborative Experts
- Scene Text Detection via Holistic, Multi-Channel Prediction
- MASTER: Multi-Aspect Non-local Network for Scene Text Recognition
- Synthetic Data for Text Localisation in Natural Images
- A Survey on Data Collection for Machine Learning: a Big Data -- AI Integration Perspective
- Semantic Jitter: Dense Supervision for Visual Comparisons via Synthetic Images
- Universal representations:The missing link between faces, text, planktons, and cat breeds
- Recent Advances in Object Detection in the Age of Deep Convolutional Neural Networks
- Scene Text Detection and Recognition: The Deep Learning Era
- Deep Structured Output Learning for Unconstrained Text Recognition
- Robust Scene Text Recognition with Automatic Rectification
- Flowing ConvNets for Human Pose Estimation in Videos
- Are Deepfakes Concerning? Analyzing Conversations of Deepfakes on Reddit and Exploring Societal Implications
- What Is Wrong With Scene Text Recognition Model Comparisons? Dataset and Model Analysis
- Scene Text Recognition with Sliding Convolutional Character Models
- Recursive Recurrent Nets with Attention Modeling for OCR in the Wild
- An End-to-End Trainable Neural Network for Image-based Sequence Recognition and Its Application to Scene Text Recognition
- Spatial Transformer Networks
- Convolutional Neural Networks with Gated Recurrent Connections
- Towards Accurate Scene Text Recognition with Semantic Reasoning Networks
- Reading Scene Text with Attention Convolutional Sequence Modeling
- CLIP4STR: A Simple Baseline for Scene Text Recognition with Pre-trained Vision-Language Model
- 2D Attentional Irregular Scene Text Recognizer
- UnrealText: Synthesizing Realistic Scene Text Images from the Unreal World
- Text Recognition in the Wild: A Survey
- TextSR: Content-Aware Text Super-Resolution Guided by Recognition
- Mask TextSpotter: An End-to-End Trainable Neural Network for Spotting Text with Arbitrary Shapes
- DAIS: Automatic Channel Pruning via Differentiable Annealing Indicator Search
- WordSup: Exploiting Word Annotations for Character based Text Detection
- Reading Scene Text in Deep Convolutional Sequences
- Beyond Visual Semantics: Exploring the Role of Scene Text in Image Understanding
- Adversarial Generation of Training Examples: Applications to Moving Vehicle License Plate Recognition
- Scene Text Recognition from Two-Dimensional Perspective
- Scene Text Synthesis for Efficient and Effective Deep Network Training
- Edit Probability for Scene Text Recognition
- ESIR: End-to-end Scene Text Recognition via Iterative Image Rectification
- 2D-CTC for Scene Text Recognition
- Reinterpreting CTC training as iterative fitting
- Exploiting Local Structures with the Kronecker Layer in Convolutional Networks
- A Survey of Deep Learning Approaches for OCR and Document Understanding
- Text Detection and Recognition in the Wild: A Review
- Learning Deep Object Detectors from 3D Models
- FACLSTM: ConvLSTM with Focused Attention for Scene Text Recognition
- Towards Unconstrained End-to-End Text Spotting
- End-to-End Subtitle Detection and Recognition for Videos in East Asian Languages via CNN Ensemble with Near-Human-Level Performance
- A pooling based scene text proposal technique for scene text reading in the wild
- Confidence Prediction for Lexicon-Free OCR
- Read Like Humans: Autonomous, Bidirectional and Iterative Language Modeling for Scene Text Recognition
- Rethinking Text Line Recognition Models
- NRTR: A No-Recurrence Sequence-to-Sequence Model For Scene Text Recognition
- SCAN: Sliding Convolutional Attention Network for Scene Text Recognition
- Fast and Robust Multiple ColorChecker Detection using Deep Convolutional Neural Networks
- A CNN Based Scene Chinese Text Recognition Algorithm With Synthetic Data Engine
- Efficient Parameter-free Clustering Using First Neighbor Relations
- CDistNet: Perceiving Multi-Domain Character Distance for Robust Text Recognition
- Transfer Learning for Scene Text Recognition in Indian Languages
- Focus-Enhanced Scene Text Recognition with Deformable Convolutions
- ResearchDoom and CocoDoom: Learning Computer Vision with Games
- RobustScanner: Dynamically Enhancing Positional Clues for Robust Text Recognition
- TextScanner: Reading Characters in Order for Robust Scene Text Recognition
- A Binary Convolutional Encoder-decoder Network for Real-time Natural Scene Text Processing
- A Multi-Object Rectified Attention Network for Scene Text Recognition
- Can Synthetic Faces Undo the Damage of Dataset Bias to Face Recognition and Facial Landmark Detection?
- Supervised mid-level features for word image representation
- WordFence: Text Detection in Natural Images with Border Awareness
- Verisimilar Image Synthesis for Accurate Detection and Recognition of Texts in Scenes
- Visual-Semantic Transformer for Scene Text Recognition
- The End-of-End-to-End: A Video Understanding Pentathlon Challenge (2020)
- A Survey on Deep Domain Adaptation and Tiny Object Detection Challenges, Techniques and Datasets
- TRIG: Transformer-Based Text Recognizer with Initial Embedding Guidance
- Recurrent Calibration Network for Irregular Text Recognition
- A Hardware-Oriented and Memory-Efficient Method for CTC Decoding
- Rethinking Irregular Scene Text Recognition
- Adaptive Adversarial Attack on Scene Text Recognition
- From Two to One: A New Scene Text Recognizer with Visual Language Modeling Network
- Frankenstein: Learning Deep Face Representations using Small Data
- Decoupled Attention Network for Text Recognition
- Generating Text Sequence Images for Recognition
- Double Supervised Network with Attention Mechanism for Scene Text Recognition
- FOTS: Fast Oriented Text Spotting with a Unified Network
- Spatial Fusion GAN for Image Synthesis
- Scene Text Image Super-Resolution in the Wild
- KISS: Keeping It Simple for Scene Text Recognition
- Gaussian Constrained Attention Network for Scene Text Recognition
- AdaDNNs: Adaptive Ensemble of Deep Neural Networks for Scene Text Recognition
- Pixel Level Data Augmentation for Semantic Image Segmentation using Generative Adversarial Networks
- Towards End-to-End Text Spotting in Natural Scenes
- Smart Library: Identifying Books in a Library using Richly Supervised Deep Scene Text Reading
- SCATTER: Selective Context Attentional Scene Text Recognizer
- DADA: Deep Adversarial Data Augmentation for Extremely Low Data Regime Classification
- Towards Boosting the Accuracy of Non-Latin Scene Text Recognition
- You Only Recognize Once: Towards Fast Video Text Spotting
- LEWIS: Latent Embeddings for Word Images and their Semantics
- SynthText3D: Synthesizing Scene Text Images from 3D Virtual Worlds
- Interpreting Neural Networks Using Flip Points
- What If We Only Use Real Datasets for Scene Text Recognition? Toward Scene Text Recognition With Fewer Labels
- Hamming OCR: A Locality Sensitive Hashing Neural Network for Scene Text Recognition
- Avoiding overfitting of multilayer perceptrons by training derivatives
- GA-DAN: Geometry-Aware Domain Adaptation Network for Scene Text Detection and Recognition
- Improving Performance of End-to-End ASR on Numeric Sequences
- Navigation Agents for the Visually Impaired: A Sidewalk Simulator and Experiments
- Human-Expert-Level Brain Tumor Detection Using Deep Learning with Data Distillation and Augmentation
- SAFL: A Self-Attention Scene Text Recognizer with Focal Loss
- STRIDE : Scene Text Recognition In-Device
- Primitive Representation Learning for Scene Text Recognition
- Revisiting Classification Perspective on Scene Text Recognition
- Deep Neural Network for Semantic-based Text Recognition in Images
- HWNet v2: An Efficient Word Image Representation for Handwritten Documents
- Self-supervised Tumor Segmentation through Layer Decomposition
- Learning to Read by Spelling: Towards Unsupervised Text Recognition
- Fused Text Recogniser and Deep Embeddings Improve Word Recognition and Retrieval
- On-Device Text Image Super Resolution
- Traditional Chinese Synthetic Datasets Verified with Labeled Data for Scene Text Recognition
- Joint Visual Semantic Reasoning: Multi-Stage Decoder for Text Recognition
- DropSample: A New Training Method to Enhance Deep Convolutional Neural Networks for Large-Scale Unconstrained Handwritten Chinese Character Recognition
- Why You Should Try the Real Data for the Scene Text Recognition
- CMFN: Cross-Modal Fusion Network for Irregular Scene Text Recognition
- AutoSTR: Efficient Backbone Search for Scene Text Recognition
- EATEN: Entity-aware Attention for Single Shot Visual Text Extraction
- Refined Gate: A Simple and Effective Gating Mechanism for Recurrent Units
- Utilizing synthetic training data for the supervised classification of rat ultrasonic vocalizations
- AON: Towards Arbitrarily-Oriented Text Recognition
- Tracking e-cigarette warning label compliance on Instagram with deep learning
- SwapText: Image Based Texts Transfer in Scenes
- EASTER: Efficient and Scalable Text Recognizer
- Context Augmentation for Convolutional Neural Networks
- A New Perspective for Flexible Feature Gathering in Scene Text Recognition Via Character Anchor Pooling
- SPIN: Structure-Preserving Inner Offset Network for Scene Text Recognition
- FedOCR: Communication-Efficient Federated Learning for Scene Text Recognition
- Matching Handwritten Document Images
- Multi-modal Text Recognition Networks: Interactive Enhancements between Visual and Semantic Features
- A study of the effect of the illumination model on the generation of synthetic training datasets
- I2C2W: Image-to-Character-to-Word Transformers for Accurate Scene Text Recognition
- Representation and Correlation Enhanced Encoder-Decoder Framework for Scene Text Recognition
- Hierarchy Composition GAN for High-fidelity Image Synthesis
- Oracle Teacher: Leveraging Target Information for Better Knowledge Distillation of CTC Models
- OmniPrint: A Configurable Printed Character Synthesizer
- Towards the Unseen: Iterative Text Recognition by Distilling from Errors
- Text is Text, No Matter What: Unifying Text Recognition using Knowledge Distillation
- Rethinking Text Segmentation: A Novel Dataset and A Text-Specific Refinement Approach
- Inductive Visual Localisation: Factorised Training for Superior Generalisation
- An optimized system to solve text-based CAPTCHA
- Parallel Scale-wise Attention Network for Effective Scene Text Recognition
- On Vocabulary Reliance in Scene Text Recognition
- Large Scale Font Independent Urdu Text Recognition System
- Integrating Scene Text and Visual Appearance for Fine-Grained Image Classification
- Similarity-based Text Recognition by Deeply Supervised Siamese Network
- Towards Fully Automated Manga Translation
- Nebula: F0 Estimation and Voicing Detection by Modeling the Statistical Properties of Feature Extractors
- ReADS: A Rectified Attentional Double Supervised Network for Scene Text Recognition
- Utilizing Resource-Rich Language Datasets for End-to-End Scene Text Recognition in Resource-Poor Languages
- Watermark retrieval from 3D printed objects via synthetic data training
- Rudder: A Cross Lingual Video and Text Retrieval Dataset
- Robust Handwriting Recognition with Limited and Noisy Data
- Adaptive Text Recognition through Visual Matching
- Auto-ML Deep Learning for Rashi Scripts OCR
- Sequence to sequence learning for unconstrained scene text recognition
- Single Shot Scene Text Retrieval
- End-to-End Interpretation of the French Street Name Signs Dataset
- Adaptive Embedding Gate for Attention-Based Scene Text Recognition
- Scene Text Recognition With Finer Grid Rectification
- Reciprocal Feature Learning via Explicit and Implicit Tasks in Scene Text Recognition
- A Machine Learning Framework for Data Ingestion in Document Images
- SAFE: Scale Aware Feature Encoder for Scene Text Recognition
- IFR: Iterative Fusion Based Recognizer For Low Quality Scene Text Recognition
- Meta Self-Learning for Multi-Source Domain Adaptation: A Benchmark
- Scene Text recognition with Full Normalization
- Distantly Supervised Semantic Text Detection and Recognition for Broadcast Sports Videos Understanding
- How many labeled license plates are needed?
- Improving Word Recognition using Multiple Hypotheses and Deep Embeddings
- Semantic Relatedness Based Re-ranker for Text Spotting
- MailLeak: Obfuscation-Robust Character Extraction Using Transfer Learning
- Implicit Feature Alignment: Learn to Convert Text Recognizer to Text Spotter
- Forward Amortized Inference for Likelihood-Free Variational Marginalization
- SmartTennisTV: Automatic indexing of tennis videos
- Extracting textual overlays from social media videos using neural networks
- Labeled Data Generation with Inexact Supervision
- Portmanteauing Features for Scene Text Recognition
- Utilizing High-level Visual Feature for Indoor Shopping Mall Navigation
- Text Recognition in Real Scenarios with a Few Labeled Samples
- It's About Time: Analog Clock Reading in the Wild
- Learning to Infer User Interface Attributes from Images