LIBER 39TH ANNUAL CONFERENCE - Statsbiblioteket

sEssIon 3.3
cLEmEns nEudEckER, asaf
tzadok (ImPact PRoJEct,
thE nEthERLands): usER
coLLaBoRatIon foR ImPRovIng
accEss to hIstoRIcaL tEXts
The paper will address how web based
collaboration tools can engage users in the
building of historical printed text resources
created by mass digitisation projects. The
drivers for developing such tools, identifying
the benefits that can be derived for both
the user community and cultural heritage
institutions, will be presented. The perceived
risks, such as new errors introduced by
the users, and the limitations of engaging
with users in this way will be set out with
the lessons that can be learnt from existing
activities, such as the National Library
of Australia’s newspaper website which
supports collaborative correction of Optical
Character Recognition (OCR) output.
The paper will present the work of the
IMPACT (Improving Access to Text,
http://www.impact-project.eu) project, a
large-scale integrating project funded by
the European Commission as part of the
Seventh Framework Programme (FP7). One
of the aims of the project is to develop tools
that help improve OCR results for historical
printed texts, specifically those works published
before the industrial production of
books from the middle of the 19th century.
The coordinator of the IMPACT project is
the KB – National library of the Netherlands.
The KB will work intensively in the
coming years to realise a digital library
that is accessible to everyone with an
Internet connection. As national library the
KB collects and maintains all publications
that appear in the Netherlands, as well
as a part of the international publications
about the Netherlands. One of the large,
labour-intensive challenges is to digitise all
the books, periodicals and newspapers that
have appeared in the Netherlands. The KB
aims to have 10% of all Dutch books, newspapers
and periodicals digitised in 2013
(60 million pages by the KB, 13 million by
third parties), as well as to offer the full-text
collections in such a way that they can be
used immediately by researchers.
To realise this goal, technological improvements
to image processing and OCR engine
technology are key. However, engaging
the user community also has an important
role to play. Utilising the intended user
can aid in achieving the levels of accuracy
currently found in born digital materials.
Improving OCR results to this level is key
to producing resources that support better
resource discovery and enabling greater
performance when applying text mining
and accessibility tools to the extracted text.
The IMPACT project will specifically develop
a tool that supports collaborative correction
and validation of OCR results and a tool to
allow user involvement in building historical
dictionaries which can be used to validate
word recognition. The technologies use the
characteristics of human perception as a
basis for error detection.
clemens neudecker holds a M.A. in
Philosophy, Computer Science and Political
Science. He has been a member of the
Munich Digitisation Centre (MDZ) from
2003-2009 and has been mostly involved
with OCR processing, authority files and
databases. He has in depth knowledge of
all steps of an in-house digitisation process,
from capture approach to online publication,
thanks to numerous responsibilities in
almost 20 digitisation projects from 2003
onwards. He currently works as Interoperability
Manager for IMPACT at the KB
National library of the Netherlands.
30 JunE 2010
43