Historical Newspapers Ground Truth

This dataset contains 50 pages of ground truth data for digitized historical newspapers from the Berlin State Library for training and evaluation of OCR/OLR systems as produced in the context of the EU ICT-PSP project Europeana Newspapers (http://www.europeana-newspapers.eu/).The dataset comprises of the following resources:gt_page.zip Ground Truth files in PAGE-XML format (cf. https://github.com/PRImA-Research-Lab/PAGE-XML)img_full.zip Full resolution scanned images in TIF formatimg_bin.zip Binarized (using the Gatos method) images in TIF formatocr_full.zip OCR (FineReaderEngine11) results for full resolution imagesocr_bin.zip OCR (FineReaderEngine11) results for binarized images