TongGuOCR: A Layout-Aware and Token-Augmented OCR Framework for Chinese Historical Documents

2026-08-08

AI Summary

This study addresses the low transcription accuracy of existing OCR systems on classical Chinese texts, which stems from their complex layouts, rare characters, and non-linear reading orders—hindering full-text search and computational analysis. To overcome these challenges, the authors propose TongGuOCR, a novel framework that integrates layout-aware preprocessing, character-level lexicon expansion, and interline spatial transition tokens to effectively model rare glyphs and intricate reading paths without requiring precise coordinate annotations. Evaluated on benchmarks such as M5HisDoc, TongGuOCR achieves state-of-the-art performance, attaining a character-level accuracy rate (AR) of 93.76%, reducing the normalized edit distance (NED) from 10.43 to 6.15, and lowering the reading-order edit distance (RO-ED) from 7.53 to 3.49, outperforming both conventional OCR systems and multimodal large language models.

Abstract

Chinese historical documents preserve valuable cultural heritage, but many collections remain accessible only as scanned page images, preventing full-text retrieval, collation, and computational analysis. Optical character recognition (OCR) can bridge this gap, but accurate transcription remains challenging because historical documents often contain complex layouts, rare characters, and nontrivial reading orders. We propose TongGuOCR, a layout-aware and token-augmented OCR framework for Chinese historical documents. First, a Layout-Aware Preprocessing module constructs and refines locally coherent recognition blocks to preserve local context while reducing interference across regions. Second, a Token-Augmented Recognition module augments the transcription target at two complementary levels: character-level vocabulary expansion gives each rare glyph a direct one-token representation and shortens its decoding path, while line-to-line transition modeling injects discrete spatial displacement tokens that guide the decoder along complex reading paths without requiring precise coordinates. Experiments on two Chinese historical document OCR benchmarks show that TongGuOCR outperforms representative traditional task-specific OCR models, general-purpose multimodal large language models (MLLMs), and OCR-oriented MLLMs. On the more challenging M5HisDoc benchmark, TongGuOCR achieves 93.76 AR and reduces NED from 10.43 to 6.15 and RO-ED from 7.53 to 3.49 relative to the best competing score for each metric. An online demo is available at https://jzzh2004.github.io/TongGuOCR.

Problem

• Chinese historical documents
• optical character recognition
• complex layouts
• rare characters
• reading order

Innovation

• layout-aware OCR
• token-augmented recognition
• Chinese historical documents
• rare character handling
• reading order modeling