@inproceedings{tuo-etal-2026-messtoclean,
    title = "{M}ess{T}o{C}lean: Evidence-Grounded Structure-Preserving Reconstruction for Real-World Degraded Exam Paper Images",
    author = "Tuo, Jiayi  and
      Tang, Cheng  and
      Wang, Zihan  and
      Zhou, Chenyue  and
      Li, Yao  and
      Ma, Yanbiao  and
      Wang, Chao  and
      Dai, Wei  and
      Wang, Mingxuan  and
      Qin, Shitong  and
      Zhao, Ziwei",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Proceedings of the 64th Annual Meeting of the {A}ssociation for {C}omputational {L}inguistics (Volume 1: Long Papers)",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl/2026.acl-long.1871/",
    pages = "40304--40322",
    ISBN = "979-8-89176-390-6",
    abstract = "Intelligent education systems often collect exam sheets as in-the-wild photos. These photos often suffer from distortions and noise caused by handwriting and occlusions, collectively referred to as Real-World Degraded Exam Images (RDEI). Structure-preserving reconstruction is key to converting RDEI into structured assets for downstream educational applications. Existing Multimodal Large Language Models (MLLMs) often fail under RDEI, leading to disrupted structure and evidence-unsupported hallucinations. To tackle these challenges, we propose MessToClean, a backbone-agnostic, evidence-driven pipeline that treats off-the-shelf MLLMs as interchangeable components. By grounding extraction in pixel-aligned evidence and enforcing post-hoc consistency auditing on recovered structures, MessToClean mitigates unsupported hallucinations and enhances both controllability and structural fidelity in question-level reconstruction. We curate RDEI-Exam from our educational platforms and evaluate across 12 state-of-the-art MLLM backbones. Across these, MessToClean improves stem consistency by 1.01-3.18{\%}, figure consistency by 0.50-49.16{\%}, and refusal F1 by 1.06-10.88{\%} across question types."
}