CULtural DImensions of deep LEarning
July 18, 2018 - July 17, 2021
NCSR budget: 193,480€
This project will develop an integrated expert system for capturing documents, enhancing images and understanding cultural documents. The capturing of an image as it is produced by a low cost scanner as well as pre-processing it, is the first stage of document processing aimed at improving scanned image quality and creating an excellent digital copy. In addition, improving the quality of a document image is a very important step because it helps to succeed in downstream processing such as fragmentation and recognition and is necessary due to the particular problems of quality and appearance of historical documents to be processed. Recognition and partition are key stages of document understanding and will assist a human expert during a cataloguing phase. The designation and identification of repeating elements such as: drop letters, dedicatory notes, typographic jewelry, titles, title decorations, character families, tables of contents and indices as well as locating the position of specific keywords that characterize the content of the documents or part of it will be automated. This will achieve a “world’s” unique application (software) that assists in the scientific documentation of cultural documents.