+ Help us process tables in Transkribus!

Information laid out in tables often seems very neat to the human eye but computers can struggle to process the tables that appear commonly in historical documents.  At READ, we are working hard to improve the automated layout analysis and recognition of tabular data.  And we need your help!  Can you share images of tables (like the below) with us so we can use them in our experiments?  Scroll down for more information…

READ researchers at the Computer Vision Lab at the Technical University of Vienna, Naver Labs Europe and the Passau Diocesan Archives would like to develop a Table Understanding module that will help us to recognise and process tables.  They would like to build up an open data set of images of historical documents containing tables.  This collection will be used to design and evaluate READ tools and as a basis for future research competitions where other computer scientists can contribute their own approaches.

If you would like to contribute to this table collection…

• Please provide at least 10 images of different tables from your documents.
• If you have a collection where the same table type is used repeatedly, please provide 25-50 pages from that collection.
• The images can include handwritten or printed text.
• Any images provided must be free to use.
• Send your images by WeTransfer or a similar file sharing site.

To send your images or ask for more information, contact:

• Hervé Déjean at Naver Labs Europe (herve.dejean@naverlabs.com)
• Eva Lang at Passau Diocesan Archives (Eva.Lang@bistum-passau.de)
• Florian Kleber at the Technical University of Vienna (kleber@cvl.tuwien.ac.at)

Thank you in advance for your help!