@inproceedings{liang-liu-2026-multi,
title = "A Multi-Stage System for {A}ncient {C}hinese {OCR} and Layout Understanding in the {E}va{H}an2026 Shared Task",
author = "Liang, KeYan and
Liu, Meiling",
editor = "Sprugnoli, Rachele and
Passarotti, Marco",
booktitle = "Proceedings of the Fourth Workshop on Language Technologies for Historical and Ancient Languages ({LT}4{HALA} 2026) @ {LREC} 2026",
month = may,
year = "2026",
address = "Palma, Mallorca (Spain)",
publisher = "ELRA Language Resources Association (ELRA)",
url = "https://aclanthology.org/2026.lt4hala-1.24/",
doi = "10.63317/4tmmz89iawet",
pages = "263--267",
abstract = "This paper presents a multi-stage system for the EvaHan2026 shared task, addressing the complex challenges of ancient Chinese optical character recognition (OCR) and layout understanding. For text recognition (Tasks A and C), we adopt parameter-efficient LoRA fine-tuning on the Qwen2.5-VL-7B-Instruct vision-language model (VLM). By directly processing full-resolution long-column images, we preserve critical spatial and contextual integrity without heuristic region cropping. For document layout analysis (Task B), we propose a novel hybrid perception-reasoning paradigm. Instead of relying solely on scaling visual detectors, we decouple localization and understanding: utilizing a YOLO-based ensemble for precise spatial bounding, and casting the VLM as a semantic verifier to eliminate spurious detections. Evaluated on the official unseen test set, our system achieves substantial improvements over the provided baselines, obtaining a 0.0441 Character Error Rate (CER) for printed OCR, a 0.0793 CER for handwritten OCR (including variants), and a 0.5118 mAP@[0.5:0.95] for layout detection. These results demonstrate that integrating VLM-based semantic reasoning into traditional visual detection pipelines is highly effective for multimodal historical document analysis."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="liang-liu-2026-multi">
<titleInfo>
<title>A Multi-Stage System for Ancient Chinese OCR and Layout Understanding in the EvaHan2026 Shared Task</title>
</titleInfo>
<name type="personal">
<namePart type="given">KeYan</namePart>
<namePart type="family">Liang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Meiling</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-05</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the Fourth Workshop on Language Technologies for Historical and Ancient Languages (LT4HALA 2026) @ LREC 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Rachele</namePart>
<namePart type="family">Sprugnoli</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Marco</namePart>
<namePart type="family">Passarotti</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>ELRA Language Resources Association (ELRA)</publisher>
<place>
<placeTerm type="text">Palma, Mallorca (Spain)</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>This paper presents a multi-stage system for the EvaHan2026 shared task, addressing the complex challenges of ancient Chinese optical character recognition (OCR) and layout understanding. For text recognition (Tasks A and C), we adopt parameter-efficient LoRA fine-tuning on the Qwen2.5-VL-7B-Instruct vision-language model (VLM). By directly processing full-resolution long-column images, we preserve critical spatial and contextual integrity without heuristic region cropping. For document layout analysis (Task B), we propose a novel hybrid perception-reasoning paradigm. Instead of relying solely on scaling visual detectors, we decouple localization and understanding: utilizing a YOLO-based ensemble for precise spatial bounding, and casting the VLM as a semantic verifier to eliminate spurious detections. Evaluated on the official unseen test set, our system achieves substantial improvements over the provided baselines, obtaining a 0.0441 Character Error Rate (CER) for printed OCR, a 0.0793 CER for handwritten OCR (including variants), and a 0.5118 mAP@[0.5:0.95] for layout detection. These results demonstrate that integrating VLM-based semantic reasoning into traditional visual detection pipelines is highly effective for multimodal historical document analysis.</abstract>
<identifier type="citekey">liang-liu-2026-multi</identifier>
<identifier type="doi">10.63317/4tmmz89iawet</identifier>
<location>
<url>https://aclanthology.org/2026.lt4hala-1.24/</url>
</location>
<part>
<date>2026-05</date>
<extent unit="page">
<start>263</start>
<end>267</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T A Multi-Stage System for Ancient Chinese OCR and Layout Understanding in the EvaHan2026 Shared Task
%A Liang, KeYan
%A Liu, Meiling
%Y Sprugnoli, Rachele
%Y Passarotti, Marco
%S Proceedings of the Fourth Workshop on Language Technologies for Historical and Ancient Languages (LT4HALA 2026) @ LREC 2026
%D 2026
%8 May
%I ELRA Language Resources Association (ELRA)
%C Palma, Mallorca (Spain)
%F liang-liu-2026-multi
%X This paper presents a multi-stage system for the EvaHan2026 shared task, addressing the complex challenges of ancient Chinese optical character recognition (OCR) and layout understanding. For text recognition (Tasks A and C), we adopt parameter-efficient LoRA fine-tuning on the Qwen2.5-VL-7B-Instruct vision-language model (VLM). By directly processing full-resolution long-column images, we preserve critical spatial and contextual integrity without heuristic region cropping. For document layout analysis (Task B), we propose a novel hybrid perception-reasoning paradigm. Instead of relying solely on scaling visual detectors, we decouple localization and understanding: utilizing a YOLO-based ensemble for precise spatial bounding, and casting the VLM as a semantic verifier to eliminate spurious detections. Evaluated on the official unseen test set, our system achieves substantial improvements over the provided baselines, obtaining a 0.0441 Character Error Rate (CER) for printed OCR, a 0.0793 CER for handwritten OCR (including variants), and a 0.5118 mAP@[0.5:0.95] for layout detection. These results demonstrate that integrating VLM-based semantic reasoning into traditional visual detection pipelines is highly effective for multimodal historical document analysis.
%R 10.63317/4tmmz89iawet
%U https://aclanthology.org/2026.lt4hala-1.24/
%U https://doi.org/10.63317/4tmmz89iawet
%P 263-267
Markdown (Informal)
[A Multi-Stage System for Ancient Chinese OCR and Layout Understanding in the EvaHan2026 Shared Task](https://aclanthology.org/2026.lt4hala-1.24/) (Liang & Liu, LT4HALA 2026)
ACL