Friedrich-Alexander-Universität
Friedrich-Alexander-Universität
 
LMU-Logo

Materials and Software

Software

Below github.com/VChristlein/puhma we provide tools which were used to analyze medieval papal charters. However, they might also be used to analyze other document images.

Provided Tools

  • Annotation tool to annotate documents, the annotations are provided in XML format

    • Box and polygon annotations
    • Extraction of annotated snippets
    • Automatic word recognition for a faster annotation

  • Snippet extraction tool to extract snippets in a more specific way than it is possible with the annotation tool,
  • Layout statistics tool to compute statistics from the XML files created by the annotation tool,
  • Plot tool to visualize statistics obtained by the layout stats tool,
  • Score-tool to score image material

Code for writer identification can be found below github.com/VChristlein/wi_wacv14 which was used for the Submission to the IEEE Conference on Winter Conference on Application of Computer Vision 2014.

 

Material

Statistics

Snippets

During the annotation process a lot of snippets, i.e. image excertps were annotated. They can be used e.g. for graph / glyph recognition or as training for various purposes.

All excerpts are from the Papstarchiv Göttingen, the full papal charters can be retreived at Opens external link in new windowwww.papsturkunden.de .

Naming convention, explained by an example: "victor_iv__14480_11630907_apostolicae_sedis_consuetudo_ph_goe__211580__graph_p.png" tells you the name of the papcy (Victor IV), the Jaffé-Nr (14480), the probable date of the appearance (11630907 = 7th of September, 1163), and last the label (p) of the instance (graph).

 

Please note that mis-labelled data may occur, we are not responsible for any possible harm.

Single

Packages

  • letters, graphs, graphical symbols, abbreviations, ligatures, words, example lines: Initiates file downloadpuhma_snippets.zip (1119,6 MB)
  • all text areas (appreciato, datum lines, first lines, cardinals signatures, middle bands, pope names, pope subscriptions, plicas, rota inscriptions, subscriptions): Initiates file downloadpuhma_snippets_textbereich.zip (620,0 MB)