Enhancing OCR Performance through Post-OCR Models: Adopting Glyph Embedding for Improved Correction
Not on the map. Why?
Abstract (arXiv)
The study investigates the potential of post-OCR models to overcome limitations in OCR models and explores the impact of incorporating glyph embedding on post-OCR correction performance. In this study, we have developed our own post-OCR correction model. The novelty of our approach lies in embedding the OCR output using CharBERT and our unique embedding technique, capturing the visual characteristics of characters. Our findings show that post-OCR correction effectively addresses deficiencies in inferior OCR models, and glyph embedding enables the model to achieve superior results, including the ability to correct individual words.
Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Builds on 15 of its 20 references that are in the map. First posted on arXiv 29 Aug 2023.
Builds on
15 of 20 references.
References · 15
- 2017Attention Is All You Needcited by 59,932
- 2015Deep Residual Learning for Image Recognitioncited by 51,809
- 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
- 2020Language Models are Few-Shot Learnerscited by 32,199
- 2023MagicBrush: A Manually Annotated Dataset for Instruction-Guided Image Editingcited by 615
- 2015An End-to-End Trainable Neural Network for Image-based Sequence Recognition and Its Application to Scene Text Recognitioncited by 519
- 2021TrOCR: Transformer-based Optical Character Recognition with Pre-trained Modelscited by 278
- 2019Character Region Awareness for Text Detectioncited by 273
- 2015Character-Aware Neural Language Modelscited by 129
- 2019Graph Convolution for Multimodal Information Extraction from Visually Rich Documentscited by 88
- 2020CharBERT: Character-aware Pre-trained Language Modelcited by 63
- 2022Human Evaluation of Text-to-Image Models on a Multi-Task Benchmarkcited by 41
3 more
Built on by
Citing papers per year: 2024: 1; 2025: 1.
2 papers.