2012 ImageCLEF WEBUPV Collection

This document describes the WEBUPV dataset compiled for the ImageCLEF
2012 Scalable image annotation task. The data mentioned here
indicates what is ready for download. However, upon request or
depending on feedback from the participants, additional data can be
released. For debugging purposes, thumbnails of the images in the
dataset can be obtained from a web server using '{IID}' the image
identifier:

  http://risenet.prhlt.upv.es/db/img/{IID}.jpg

The following is the directory structure of the collection, and bellow
there is a brief description of what each compressed file
contains. The corresponding MD5 checksums of the files shown (for
verifying a correct download) can be found in the md5sums.txt.

Directory structure
-------------------

.
|
|--- README.txt
|--- md5sums.txt
|--- webupv_train_lists.zip
|--- webupv_devel_lists.zip
|--- webupv_test_lists.zip
|--- baseline.zip
|
|--- feats_textual/
|      |
|      |--- webupv_train_textual.rawfeat.gz
|      |--- webupv_train_textual.scofeat.gz
|      |--- webupv_train_textual.keywords.gz
|
|--- feats_visual/
       |
       |--- webupv_{train|devel|test}_visual_gist.feat.gz
       |--- webupv_{train|devel|test}_visual_sift_*.feat.gz
       |--- webupv_{train|devel|test}_visual_csift_*.feat.gz
       |--- webupv_{train|devel|test}_visual_rgbsift_*.feat.gz
       |--- webupv_{train|devel|test}_visual_opponentsift_*.feat.gz
       |--- webupv_{train|devel|test}_visual_colorhist.feat.gz

Contents of files
-----------------

* webupv_train_lists.zip
  -> train_iids.txt : IDs of the images in the training set (250000).
  -> train_rids.txt : IDs of the webpages in the training set.
  -> train_rimgsrc.txt : The URLs of the images as referenced in each
                         of the webpages. This can also be useful as a
                         textual feature.

* webupv_devel_lists.zip
  -> devel_iids.txt : IDs of the images in the development set (1000).
  -> devel_concepts.txt : List concepts for the development set.
  -> devel_gnd.txt : Ground truth concepts for the development set
                     images.

* webupv_test_lists.zip
  -> test_iids.txt : IDs of the images in the test set (2000).
  -> test_concepts.txt : List concepts for the test set.
  -> test_gnd.txt : Ground truth concepts for the test set images.

* baseline.zip

  An archive that includes code for computing the evaluation measures
  for two baseline techniques for the "Scalable concept image
  annotation" subtask. See the included README.txt for details.

* feats_textual/webupv_train_textual.rawfeat.gz

  The raw text extracted from the webpages near where the images
  appeared. Each line starts with the image and webpage IDs followed
  by the text extracted. The position of the image within the text is
  indicated by the special word '{X}'. The extracted text is somewhat
  filtered (e.g. there are no HTML tags), although removed words and
  tags have been replaced by full stops '.' to preserve word
  distances. The title of the webpage is always included, and it is
  the first sentence of the text. In total the file has 275749 lines
  since the images can appear in more than one webpage.

* feats_textual/webupv_train_textual.scofeat.gz

  The processed text extracted from the webpages near where the images
  appeared. Each line corresponds to one image, having the same order
  as the train_iids.txt list. The lines start with the image ID,
  followed by the number of extracted unique words and the
  corresponding word-score pairs. The scores were derived taking into
  account 1) the term frequency (TF), 2) the document object model
  (DOM) attributes, and 3) the word distance to the image. The scores
  are all integers and for each image the sum of scores is always
  <=100000 (i.e. it is normalized).

* feats_textual/webupv_train_textual.keywords.gz

  The words used to find the images when querying image search
  engines. Each line corresponds to an image (in the same order as
  in train_iids.txt). The lines are composed of triplets:

    [keyword] [rank] [search_engine]

  where [keyword] is the word used to find the image, [rank] is the
  position given to the image in the query, and [search_engine] is a
  single character indicating in which search engine it was found
  ('g':google, 'b':bing, 'y':yahoo).

* feats_visual/webupv_*.feat.gz

  The visual features in a simple ASCII text sparse format. The first
  line of the file indicates the number of vectors (N) and the
  dimensionality (DIMS). Then each line corresponds to one vector,
  starting with the number of non-zero elements and followed by pairs
  of dimension-value, being the first dimension 0. In summary the file
  format is:

    N DIMS
    nz1 Dim(1,1) Val(1,1) ... Dim(1,nz1) Val(1,nz1)
    nz2 Dim(2,1) Val(2,1) ... Dim(2,nz2) Val(2,nz2)
    ...
    nzN Dim(N,1) Val(N,1) ... Dim(N,nzN) Val(N,nzN)

  The order of the features is the same as in the lists
  devel_iids.txt, test_iids.txt and train_iids.txt.

  The procedure to extract the SIFT based features in this
  subdirectory was conducted as follows. Using the ImageMagick
  software, the images were first rescaled to having a maximum of 240
  pixels, of both width and height, while preserving the original
  aspect ratio, employing the command:

    convert {IMGIN}.jpg -resize '240>x240>' {IMGOUT}.jpg

  Then the SIFT features where extracted using the ColorDescriptor
  software from Koen van de Sande
  (http://koen.me/research/colordescriptors). As configuration we
  used, 'densesampling' detector with default parameters, and a hard
  assignment codebook using a spatial pyramid as
  'pyramid-1x1-2x2'. The number in the file name indicates the size of
  the codebook. All of the vectors of the spatial pyramid are given in
  the same line, thus keeping only the first 1/5th of the dimensions
  would be like not using the spatial pyramid. The codebook was
  generated using 1.25 million randomly selected features and the
  k-means algorithm.

Contact
-------

For further questions, please contact:
  Mauricio Villegas <mauvilsa@upv.es>