Articolul precedent |
Articolul urmator |
631 2 |
Ultima descărcare din IBN: 2021-04-15 13:32 |
Căutarea după subiecte similare conform CZU |
004.912 (18) |
Прикладные информационные (компьютерные) технологии. Методы основанные на применении компьютеров (438) |
SM ISO690:2012 BURTSEVA, Lyudmila, COJOCARU, Svetlana, MALAHOV, Ludmila, COLESNICOV, Alexandru. Platform for processing heterogeneous documents. In: Patrimoniul cultural de ieri – implicații în dezvoltarea societății durabile de mâine, Ed. 3, 11-12 februarie 2021, Chişinău. Chișinău, Republica Moldova: 2021, Ediția 3, pp. 46-47. ISSN 2558 – 894X. |
EXPORT metadate: Google Scholar Crossref CERIF DataCite Dublin Core |
Patrimoniul cultural de ieri – implicații în dezvoltarea societății durabile de mâine Ediția 3, 2021 |
|||||||
Conferința "Yesterday’s heritage – implications for the development of tomorrow’s sustainable society" 3, Chişinău, Moldova, 11-12 februarie 2021 | |||||||
|
|||||||
CZU: 004.912 | |||||||
Pag. 46-47 | |||||||
|
|||||||
Descarcă PDF | |||||||
Rezumat | |||||||
Most documents (historical or contemporary, handwritten or printed) have a heterogeneous character because, along with the text itself, they also contain other elements: mathematical and chemical formulas, musical scores, diagrams, images, etc. In our research, we will operate with documents the non-textual components of which permits their digital representation in one or more scripting languages. Given the fact that a complete automation of the processing of heterogeneous documents is difficult or maybe even impossible to perform, we aim to create a platform that would combine automatic, semi-automatic and manual processing modules. Thus, we can automatically perform scanning, recognizing document segments according to their type, assembling presentations in scripting language with metadata integration, reconstructing the page image based on the script, checking within certain limits. In a semi-automated way, the image quality will be improved, the page layout will be analyzed, tasks will be distributed for manual verification. For cases where automated or semi-automated verification procedures could not be applied, it will be operated manually, the platform providing some support for this mode as well. The architecture of the platform proposed in the paper is aimed at maximizing the support of the mentioned steps by creating the “single window” for access to all processing tools and by reducing manual operations. * Acknowledgement. This paper was written as part of the re-search project 20.80009.5007.22 „Intelligent information systems forsolving ill-structured problems, processing knowledge and big data” |
|||||||
|