[Corpora-List] CFP announcement: OntoImage'2006

GREFENSTETTE Gregory 206823 gregory.grefenstette at cea.fr
Wed Jan 11 10:28:00 CET 2006


Objet : Italy: OntoImage'2006 Workshop - Call for Papers



http://www.lrec-conf.org/lrec2006/IMG/pdf/LRECworkshopOntoImage.htm
<http://www.lrec-conf.org/lrec2006/IMG/pdf/LRECworkshopOntoImage.htm>



Call for Papers

International Workshop

OntoImage'2006, Language Resources for Content-Based Image Retrieval May
22, 2006, Genoa, Italy

in conjunction with LREC'2006)



-----------------------------------------------------

Deadline for Abstract Submissions: February 20th 2006 Submission per
email to: Gregory.Grefenstette at cea.fr



-----------------------------------------------------

Background

The "semantic gap" between low-level image description (texture, shape,
color) and what people perceive in images is recognized as a difficult
and unsolved problem by all content-based image researchers. With the
growing number of images being created and stored each day, progress
must be made on automatic methods for indexing and retrieving images
based on their content. The vocabulary that would be most effective for
this indexing is poorly understood today.



A number of approaches for using text to retrieve image have been
proposed (hand assignment of descriptions [e.g. FLICKR,
GRAPHICOBSESSION]; extracting terms from captions and from text linked
to images [GOOGLE]; associating word and image characteristics [Barnard
et al., 2003; Dance et al., 2004]. Recently, ontologies have been
suggested as the proper representation for image indexing [Town &
Sinclair, 2004]. Much work remains to be done to understand the problem

>From a lexical point of view.




This half-day workshop will present up to 9 papers examining problems
concerning the lexical characteristics of the language that can or
should be used to describe image content. The goal of the workshop is to
address the following questions:



-- What vocabulary should be used to describe images?

-- How can vocabulary items be automatically associated with images

-- for indexing and retrieval?

-- What visual characteristics of images can be discovered by text

-- processing (color, size, location, shape, texture)?

-- What elements in a lexicon correspond to visualisable objects?

-- What language resources exist and what other ones are needed for

-- content-based image retrieval?

-- Can specific ontologies be created and automatically populated for

-- content-based image retrieval?

-- Can text processing identify what objects would appear against what

-- background (e.g. bird against a sky)?





------------------------------------------------------

Papers



We expect extended abstracts describing research addressing one of the
above questions of about 1000 words to be submitted as WORD, PDF or
ASCII documents per email to the following address:
Gregory.Grefenstette at cea.fr. The final papers should not have more than
4 pages.



------------------------------------------------------



Important Dates

Deadline for Abstracts February 20, 2006

Notification of acceptance March 10, 2006

Final Papers March 31, 2006

Workshop May 22, 2006 (afternoon)



------------------------------------------------------

Location



Magazzini del Cotone Conference Center, Genoa, Italy. Co-located with
LREC'2006 http://www.lrec-conf.org/lrec2006
<http://www.lrec-conf.org/lrec2006>





------------------------------------------------------

Organizers

Gregory Grefenstette, Commissariat ` l'Energie Atomique (CEA LIST),
Fontenay-aux-Roses, France



Mark Sanderson, Department of Information Studies, University of
Sheffield, Sheffield, UK



Frangoise Preteux, Institut National des Tilicommunications (ARTEMIS),
Evry, France



------------------------------------------------------



Questions



Questions about the workshop can be addressed to:
gregory.grefenstette at cea.fr



See also http://www.lrec-conf.org/lrec2006/rubrique.php3?id_rubrique=6
<http://www.lrec-conf.org/lrec2006/rubrique.php3?id_rubrique=6>







More information about the Corpora-archive mailing list