ImageCLEF

From Chorus
Jump to: navigation, search
ImageCLEF
http://imageclef.org/
Event Short Name ImageCLEF
Event Full Name Image Retrieval in CLEF (Cross-Language Evaluation Forum)
Organizer several organisers, overall coordination: MedGIFT, HES-SO, Switzerland
Domain Medical, Photography, Web, Botanical, Robot
Media Image
Task Retrieval, Classification
Date 17-20 September 2012
Place Rome, Italy
Co-located with CLEF 2012
URL http://imageclef.org/


Description

ImageCLEF aims to provide an evaluation forum for the cross–language annotation and retrieval of images. Motivated by the need to support multilingual users from a global community accessing the ever growing body of visual information, the main goal of ImageCLEF is to support the advancement of the field of visual media analysis, indexing, classification, and retrieval, by developing the necessary infrastructure for the evaluation of visual information retrieval systems operating in both monolingual, cross–language and language-independent contexts. ImageCLEF aims at providing reusable resources for such benchmarking purposes.

ImageCLEF launched in 2003 as part of CLEF with the goal is to provide support for the evaluation of 1) language-independent methods for the automatic annotation of images with concepts, 2) multimodal information retrieval methods based on the combination of visual and textual features, and 3) multilingual image retrieval methods, so as to compare the effect of retrieval of image annotations and query formulations in several languages.

To meet its objectives, ImageCLEF 2012 will organise tasks that benchmark the annotation and retrieval of diverse images such as general photographic and medical images, as well as domain-specific tasks such as plant identification and robot vision. These evaluation tasks aim to support and promote research that addresses key challenges in the field including:

  1. visual image annotation with concepts at various levels of abstraction that relies not only on manual, and thus reliable, training data, but also on automatically acquired, and thus noisy, labelled samples,
  2. modality classification and retrieval in the medical domain that moves closer to clinical practice and routine through classification tasks that consider complex, hierarchically organised, classes of modalities and retrieval tasks that support medical practitioners in their decision making,
  3. scientific multimedia data management through the particular case of botanical data identification,
  4. the shift in the area of robot vision from visual place recognition to multimodal place recognition, and
  5. the combination of visual features with highly diverse textual evidence, ranging from unstructured, to semi-structured, or even highly-structured textual annotations, as well as semantically enriched data.


Tasks

  • Photo retrieval (2003-2009)
  • Medical image retrieval (2004-2012)
  • Medical image annotation (2005-2009)
  • Photo annotation (2006-2012)
  • Wikipedia image retrieval (2008-2011)
  • Robot Vision (2009-2010, 2012)
  • Plant identification (2011-2012)


Important Dates

Each of the tasks sets its own schedule. A (tentative) global schedule for 2012 can be found below:

  • 1.3.2012: registration opens for all ImageCLEF tasks
  • 15.3.2012 - 30.4.2012: data release (depending on the task)
  • 15.4.2012 - 15.5.2012: topic release (depending on the task)
  • 15.5.2012: registration closes for all ImageCLEF tasks
  • 1.6.2012 - 30.6.2012: submission of runs (depending on the task)
  • 15.7.2012 - 31.7.2012: release of results (depending on the task)
  • 15.8.2012 : submission of working notes papers
  • 17.9.2012 - 20.9.2012: CLEF 2012 Conference, Rome, Italy


Related Resources

External Links

Personal tools
CHORUS+