HeDy: AN OPEN PLATFORM FOR DIGITIZATION OF PRINTED DOCUMENTS

Summary. The article is a synthesis of the results obtained by the authors in the field of digitization and transliteration of texts, with a focus on works printed in Romanian with Cyrillic characters. Recent advancements in digital technologies, particularly in artificial intelligence, underscore the need to automate the digitization of texts, which serve as essential resources for developing large linguistic models. The volume of scanned document archives has dramatically increased due to cultural heritage preservation projects. However, existing document processing systems often struggle with accurately displaying content containing heterogeneous components. The HeDy platform addresses this challenge by offering solutions to enhance all stages of processing heterogeneous documents. These stages include image preprocessing, optical character recognition (OCR), document segmentation, and the transliteration of Romanian texts from Cyrillic to modern Latin alphabets. The article is a synthesis of the results obtained by the authors in the field of digitization and transliteration of texts, with a focus on works printed in Romanian with Cyrillic characters.Key words: digitization, transliteration, heterogeneous documents.