LVM-OCR: A Transformer-Based Architecture for Context-Aware Document Understanding

In an era where digital transformation demands seamless document processing, we introduce a groundbreaking approach to Optical Character Recognition that harnesses the remarkable capabilities of Large Vision Models. Our innovative LVM-OCR framework represents a paradigm shift in how machines read and understand text, moving beyond traditional pattern matching to genuine visual comprehension. By weaving together transformer-based architectures with intelligent text-awareness mechanisms, we have created a system that does not just recognize characters, it understands context, adapts to complexity, and learns from ambiguity. Our extensive experiments reveal something extraordinary: a 15.3% leap in accuracy over conventional methods, with particular brilliance in deciphering degraded historical documents, multilingual texts, and complex layouts that would confound traditional systems. This is not merely an incremental improvement; it is a fundamental reimagining of what OCR technology can achieve.

• Publication year

  2025

• Venue

  2025 1st International Conference on Emerging Trends in Information Systems and Informatics (ICETISI)

• Publication date

  2025-12-01

• Fields of study

  Not labeled

• Identifiers
  DOI 10.1109/ICETISI67983.2025.11406044
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Understanding the architecture of vision transformer and its variants: A review

  2024cited by this paper
• Advancements and Challenges in Arabic Optical Character Recognition: A Comprehensive Survey

  2023cited by this paper
• A Resource-Efficient Keyword Spotting System Based on a One-Dimensional Binary Convolutional Neural Network

  2023cited by this paper
• A Comprehensive Study of Optical Character Recognition

  2022cited by this paper
• Swin Transformer: Hierarchical Vision Transformer using Shifted Windows

  2021cited by this paper
• TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models

  2021cited by this paper
• Training data-efficient image transformers & distillation through attention

  2020cited by this paper
• An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale

  2020cited by this paper
• ASTER: An Attentional Scene Text Recognizer with Flexible Rectification

  2019cited by this paper
• ICDAR2019 Robust Reading Challenge on Multi-lingual Scene Text Detection and Recognition — RRC-MLT-2019

  2019cited by this paper
• ICDAR2019 Competition on Scanned Receipt OCR and Information Extraction

  2019cited by this paper
• Show, Attend and Read: A Simple and Strong Baseline for Irregular Text Recognition

  2018cited by this paper
• Attention is All you Need

  2017cited 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
• Neural Machine Translation by Jointly Learning to Align and Translate

  2014cited by this paper
• An Overview of the Tesseract OCR Engine

  2007cited by this paper
• Novel neural modulators.

  2003cited by this paper
• The IAM-database: an English sentence database for offline handwriting recognition

  2002cited by this paper
• Twenty Years of Document Image Analysis in PAMI

  2000cited by this paper
• Historical review of OCR research and development

  1992cited by this paper

• No citing papers are available for this paper.