September 5, 2025 Dataset Open Access

Hussein Mohammed; Quang-Vinh Dang

DataCite XML Export

<?xml version='1.0' encoding='utf-8'?>
<resource xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns="http://datacite.org/schema/kernel-3" xsi:schemaLocation="http://datacite.org/schema/kernel-3 http://schema.datacite.org/meta/kernel-3/metadata.xsd">
  <identifier identifierType="DOI">10.25592/uhhfdm.18703</identifier>
  <creators>
    <creator>
      <creatorName>Hussein Mohammed</creatorName>
      <nameIdentifier nameIdentifierScheme="ORCID" schemeURI="http://orcid.org/">0000-0001-5020-3592</nameIdentifier>
      <affiliation>Universität Hamburg</affiliation>
    </creator>
    <creator>
      <creatorName>Quang-Vinh Dang</creatorName>
      <nameIdentifier nameIdentifierScheme="ORCID" schemeURI="http://orcid.org/">0000-0002-6715-7112</nameIdentifier>
      <affiliation>Universität Hamburg</affiliation>
    </creator>
  </creators>
  <titles>
    <title>The DLA-RMR dataset: Annotated subset of RMR notebooks for CVC development</title>
  </titles>
  <publisher>Universität Hamburg</publisher>
  <publicationYear>2025</publicationYear>
  <subjects>
    <subject>page detection</subject>
    <subject>word detection</subject>
    <subject>colour recognition</subject>
    <subject>recognition of writing implement</subject>
    <subject>visual navigation</subject>
    <subject>computational visual cataloguing</subject>
  </subjects>
  <dates>
    <date dateType="Issued">2025-09-05</date>
  </dates>
  <language>en</language>
  <resourceType resourceTypeGeneral="Dataset"/>
  <alternateIdentifiers>
    <alternateIdentifier alternateIdentifierType="url">https://www.fdr.uni-hamburg.de/record/18703</alternateIdentifier>
  </alternateIdentifiers>
  <relatedIdentifiers>
    <relatedIdentifier relatedIdentifierType="DOI" relationType="IsSupplementedBy">10.25592/uhhfdm.17809</relatedIdentifier>
    <relatedIdentifier relatedIdentifierType="DOI" relationType="IsSupplementedBy">10.25592/uhhfdm.17613</relatedIdentifier>
    <relatedIdentifier relatedIdentifierType="DOI" relationType="IsSupplementedBy">10.25592/uhhfdm.17615</relatedIdentifier>
    <relatedIdentifier relatedIdentifierType="DOI" relationType="IsPartOf">10.25592/uhhfdm.17931</relatedIdentifier>
  </relatedIdentifiers>
  <version>1.1</version>
  <rightsList>
    <rights rightsURI="https://creativecommons.org/licenses/by/4.0/legalcode">Creative Commons Attribution 4.0 International</rights>
    <rights rightsURI="info:eu-repo/semantics/openAccess">Open Access</rights>
  </rightsList>
  <descriptions>
    <description descriptionType="Abstract">&lt;p&gt;What&amp;rsquo;s new in this version: In this version, we added colour ground truth for the test set to permit quantitative evaluation, additional details about the images used, and a carefully selected train&amp;ndash;test split to avoid any leakage at the notebook level.&lt;br&gt;
&lt;br&gt;
This&amp;nbsp;dataset is structured into four components, each serving a distinct role in the development of a&amp;nbsp;document analysis system.&amp;nbsp;&lt;/p&gt;

&lt;ol&gt;
	&lt;li&gt;
	&lt;p&gt;&lt;strong&gt;Word-level annotations&lt;/strong&gt; are provided in the file &lt;code&gt;word_annotations_for_cropped_images.json&lt;/code&gt;. These annotations describe the images contained in the &lt;code&gt;cropped_images&lt;/code&gt; folder. Each entry specifies the location of a word as a polygon, together with its orientation (horizontal, vertical, or tilted) and the type of writing implement used (ink or pencil). Additional metadata, such as bounding boxes and segmentation areas, is also included.&lt;/p&gt;
	&lt;/li&gt;
	&lt;li&gt;
	&lt;p&gt;&lt;strong&gt;Cropped images&lt;/strong&gt; are stored in the &lt;code&gt;cropped_images&lt;/code&gt; folder. This set comprises 50 images, each containing only the primary page extracted from the corresponding full notebook scans.&lt;/p&gt;
	&lt;/li&gt;
	&lt;li&gt;
	&lt;p&gt;&lt;strong&gt;Full images&lt;/strong&gt; are located in the &lt;code&gt;full_images&lt;/code&gt; folder. This collection also contains 50 items, representing the complete notebook scans in which the primary page appears alongside other material.&lt;/p&gt;
	&lt;/li&gt;
	&lt;li&gt;
	&lt;p&gt;&lt;strong&gt;Page-level annotations&lt;/strong&gt; are contained in the &lt;code&gt;page_annotations&lt;/code&gt; folder. These are provided in YOLO format, with a single class (&lt;code&gt;page&lt;/code&gt;) defined in &lt;code&gt;classes.txt&lt;/code&gt;. Each annotation file specifies the bounding box of the primary page within the corresponding image in the &lt;code&gt;full_images&lt;/code&gt; folder.&lt;/p&gt;
	&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;Examples illustrate the annotation structure. In the JSON file, a typical word annotation records polygon coordinates, the attribute &lt;code&gt;&amp;quot;orientation&amp;quot;: &amp;quot;horizontal&amp;quot;&lt;/code&gt;, and &lt;code&gt;&amp;quot;writing_tool&amp;quot;: &amp;quot;pencil&amp;quot;&lt;/code&gt;. In the YOLO annotations, a sample entry such as &lt;code&gt;0 0.499023 0.500776 0.777344 0.816912&lt;/code&gt; denotes the normalised coordinates of the primary page bounding box.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Acknowledgement:&lt;/strong&gt;&lt;/p&gt;

&lt;p&gt;The research for this work was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under Germany&amp;rsquo;s Excellence Strategy - EXC 2176 &amp;lsquo;Understanding Written Artefacts: Material, Interaction and Transmission in Manuscript Cultures&amp;rsquo;, project no. 390893796. The research was conducted within the scope of the Centre for the Study of Manuscript Cultures (CSMC) at Universit&amp;auml;t Hamburg.&lt;br&gt;
&lt;br&gt;
The images are taken from notebook pages of Rainer Maria Rilke, from the Deutsche Literaturarchiv Marbach (DLA), A:Rilke-Archiv Gernsbach.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;We thank Hui Xu for her support in annotating the images.&lt;/p&gt;</description>
  </descriptions>
</resource>