{
    "Name": "AraMS-28k",
    "Volume": 28600.0,
    "Unit": "sentences",
    "License": "CC BY-NC-SA 4.0",
    "Link": "https://doi.org/10.5281/zenodo.22095333",
    "HF_Link": "unknown",
    "Year": 2025,
    "Source": [
        "books"
    ],
    "Form": "text",
    "Domain": [
        "history",
        "religion",
        "general"
    ],
    "Annotation_Style": [
        "LLM annotation",
        "human annotation",
        "human validation"
    ],
    "Description": "Public line-level dataset for historical Arabic HTR",
    "Provider": [
        "Higher School of Computer Science (ESI-SBA)"
    ],
    "Derived_From": [],
    "Partial": false,
    "Paper_Title": "AraMS-28k: The Largest Publicly Released Line-Level Dataset of Historical Arabic Manuscripts with Margin and Insertion-Anchor Annotations",
    "Paper_Link": "https://arxiv.org/pdf/2608.26921v1.pdf",
    "Tokenized": false,
    "Host": "HuggingFace",
    "Access": "Free",
    "Cost": "",
    "Has_Splits": true,
    "Tasks": [
        "handwriting recognition"
    ],
    "Venue_Title": "unknown",
    "Venue_Type": "preprint",
    "Venue_Name": "arXiv",
    "Authors": [
        "Mohamed Guechaoui",
        "Mohamed Diaa Zellagui",
        "Souleyman Chaib",
        "Sahraoui Dhelim"
    ],
    "Affiliations": [
        "Higher School of Computer Science (ESI-SBA)"
    ],
    "Abstract": "We introduce AraMS-28k, the largest publicly released line-level dataset of genuine historical Arabic manuscripts, comprising 14 books, 3,043 pages, and 28,600 annotated text lines. Thirteen books are hand-copied manuscripts spanning three script traditions\u2013Naskh, Ruq\u2018ah, and Maghrebi\u2013and one is a lithographed printed edition included to broaden format diversity. Each line is labelled as main-text or margin, and margin lines that have an unambiguous attachment point in the main text are further annotated with an insertion anchor, recovering the manuscript\u2019s true non-linear reading order at line-level granularity\u2013to our knowledge the first such annotation released for a historical Arabic manuscript corpus. Because reference transcriptions are fully vocalised while manuscript hands are typically undiacritised, we release both the raw diacritised transcription and a diacritic-normalised counterpart for every line. The dataset was constructed with RefLAM[1], a reference-grounded annotation pipeline that aligns multimodal-LLM OCR against independently sourced clean transcriptions and routes every line through human review, combining automatic verification with expert oversight. We describe the construction and quality-control process, present the annotation schema, report dataset statistics at both the corpus and per-book level, and provide baseline HTR results using Kraken and HATFormer, including across-script generalisation gradient from in-distribution pages to a fully unseen books. AraMS-28k is released with page images, line-level annotations, and fixed train/val/test splits under CC BY-NC-SA 4.0 to support reproducible research on Arabic manuscript recognition, layout analysis, and reading-order recovery.",
    "Dialect_Subsets": [],
    "Dialect": "Classical Arabic",
    "Language": "ar",
    "Script": "Arab",
    "Added_By": "qwen/qwen3.6-35b-a3b"
}