Time filter

Source Type

Mazurek C.,Poznan Supercomputing And Networking Center Ul Z | Mielnicki M.,Poznan Supercomputing And Networking Center Ul Z | Nowak A.,Poznan Supercomputing And Networking Center Ul Z | Stroinski M.,Poznan Supercomputing And Networking Center Ul Z | And 2 more authors.
Studies in Computational Intelligence | Year: 2013

One of the tasks undertaken in PSNC in the frame of the SYNAT project is the design of architecture for the scientific information system allowing to integrate heterogeneous distributed services necessary to build knowledge base and innovative applications based on it. This paper contains overall description of the designed architecture with the analysis of its most important technical aspects, features and possible weak points. It also presents initial results from the test deployment of the first prototype of several architecture components, including initial aggregation and processing of several millions of metadata records from several tenths of network services. © 2013 Springer-Verlag Berlin Heidelberg.


Dudczak A.,Poznan Supercomputing and Networking Center Ul Z | Kmieciak M.,Poznan Supercomputing and Networking Center Ul Z | Mazurek C.,Poznan Supercomputing and Networking Center Ul Z | Stroinski M.,Poznan Supercomputing and Networking Center Ul Z | And 2 more authors.
Studies in Computational Intelligence | Year: 2013

This paper describes improvements which can be included in digitisation workflow to increase the number and enhance the quality of full text representations of historical documents.This kind of documents is well represented in Polish digital libraries and can offer interesting opportunities for digital humanities researchers. Proposed solution focuses on changing existing approach to OCR and simplifying the manual process of text correction through crowdsourcing. Tools required to implement these enhancements are available in Virtual Transcription Laboratory (VTL) prototype, developed in the framework of SYNAT (http://www.synat.pl) project. In the last chapter paper describes results of the experiment with the custom OCR recognition profiles which proves that proposed approach is a viable alternative to existing OCR practices. © 2013 Springer-Verlag Berlin Heidelberg.

Loading Poznan Supercomputing And Networking Center Ul Z collaborators
Loading Poznan Supercomputing And Networking Center Ul Z collaborators