|
relevance judgements
|
Henning Müller
|
Oct 22, 2001 03:26 PDT
|
Dear Benchathletes,
now that the annotation tool is finished, the vocabulary more or less
fixed and the image database selected, we should start to think about
the next steps we have to take before we can present the framework at
the SPIE conference in January.
Should we make a list of all the people who are willing to participate
in the annotation effort? I think it will be a lot of work, so we should
distribute it as good as possible.
Who would manage a list of voluntary annotators? Corinne?
The step after this is to extract relevance judgements from the
annotation of the images.
Does anybody of you have experiences with this?
We could of course use text retrieval techniques to get a similarity
score of images based on the annotation, but this is may be not the best
solution.
I think that it is very important to have different sets of relevance
judgements for the same query image, to model different users opinions.
All the relevance sets we got from real users are differing for every
single user.
It is also very important to show that a retrieval system is able to
adapt to the users needs and the needs can be very different result
images for the same query image.
What do you think?
Cheers, Henning
--
----------------------------------------------------------
Henning Mueller, Computer Vision Group
Computer Science Department, University of Geneva
24, rue du General Dufour, CH-1211 Geneva 4, SWITZERLAND
Phone : +41(22)705 7633; fax: +41(22)705 7780
Henning.-@cui.unige.ch
----------------------------------------------------------
|
|
 |
|
