Skip to Content

International Comittee of the Red Cross

Switzerland ✧ 2023-2024

Project


Automatic indexing of names from lists of French prisoners of war from the Second World War.


Corpus


700,000 pages of registers

Processing 


-> Automatic Text Recognition

-> Information Extraction

Collaborative annotation campaigns


Processing workflow 


  • Annotation of 500 pages by archivists on the Arkindex Callico extension

  • Extraction of personally identifiable information using a hybrid model combining handwriting recognition (HTR) and named entity recognition (NER)