📰 Layout and Structure

The purpose of providing Ground-Truth with layout and structure information is to ensure that both in the training process and in the end result of text recognition (OCR+OLR), all regions recognizable on a page can be recognized and classified according to the requirements. The trained models should contribute to realizing the full text recognition of the prints from the 16th–18th century as a mass application and enable the text accuracy in the overall result of the text recognition process to meet the highest scientific requirements. As with the text transcription, the Ground-Truth can be available in different recognition depths (levels).