**Description:**

We provide two large-scale collections for experiments with the **libmsvm** package: 
  - the trecvid 2015 collections for semantic indexing task and 
  - the voc 2012 collections.

Each data-set is divided into the development and test data-sets.

The development data-set is also splitted into two subsets (training and validating) in-order to be used with cross-validations.
The descriptor and the labels are provided for all samples in different ways (i.e. separated files, one file, binary and .svm formats).

**[[http://mrim.imag.fr/libmsvm/data/data_exp/trecvid_sin15/|Trecvid_sin15]]**

  *   annotations of [[http://mrim.imag.fr/libmsvm/data/data_exp/trecvid_sin15//2013/cat_2013f.txt|346 visual concepts]] of Trecvid SIN 2013 are provided.
  *   [[http://mrim.imag.fr/libmsvm/data/data_exp/trecvid_sin15//2013/|Full collection 2013]]
      - [[http://mrim.imag.fr/libmsvm/data/data_exp/trecvid_sin15/2013/LIG_alex_goog_vgg_earlypw1.000p294pw0.610.bin|descriptor binary]]
      - separated annotations ' Two formats i) [[http://mrim.imag.fr/libmsvm/data/data_exp/trecvid_sin15/2013/2013f.ann|binary-labels]] (-1, 0, +1) and ii) [[http://mrim.imag.fr/libmsvm/data/data_exp/trecvid_sin15/2013/2013f_mlabel.ann|multi-labels]] {+-1 .. +-346}'
      - .SVM Format (i.e. annotation + descriptor vector in sparse mode) ' Two formats i) [[http://mrim.imag.fr/libmsvm/data/data_exp/trecvid_sin15/2013/LIG_alex_goog_vgg_earlypw1.000p294pw0.610.svm|binary-labels]] and ii) [[http://mrim.imag.fr/libmsvm/data/data_exp/trecvid_sin15/2013/LIG_alex_goog_vgg_earlypw1.000p294pw0.610_mlabel.svm|multi-labels]]'	

  * Sub collections :Each subset contains a **select.txt** file, which indicates the keylist samples to be selected from the full collection.
      -  Developement set [[http://mrim.imag.fr/libmsvm/data/data_exp/trecvid_sin15/2013d/|2013d]]: divided into two subsets [[http://mrim.imag.fr/libmsvm/data/data_exp/trecvid_sin15/2013x/|2013x]]and [[http://mrim.imag.fr/libmsvm/data/data_exp/trecvid_sin15/2013y/|2013y]]
      - Test set [[http://mrim.imag.fr/libmsvm/data/data_exp/trecvid_sin15/2013z/|2013z]].
  
**[[http://mrim.imag.fr/libmsvm/data/data_exp/voc12/|voc12]]**

  *   annotations of [[http://mrim.imag.fr/libmsvm/data/data_exp/voc12/all/cat_20.txt|20 visual concepts]] of Pascal-VOC 2012 are provided.
  *   [[http://mrim.imag.fr/libmsvm/data/data_exp/voc12/all/|Full collection 2012]]
      - [[http://mrim.imag.fr/libmsvm/data/data_exp/voc12/all/LIG_alex_goog_vgg_earlypw1.000p294pw0.610.bin|descriptor binary]]
      - separated annotations ' Two formats i) [[http://mrim.imag.fr/libmsvm/data/data_exp/voc12/all/2012.ann|binary-labels]] (-1, 0, +1) and ii) [[http://mrim.imag.fr/libmsvm/data/data_exp/voc12/all/2012_mlabel.ann|multi-labels]] {+-1 .. +-20}'
      - .SVM Format (i.e. annotation + descriptor vector in sparse mode) ' Two formats i)   [[http://mrim.imag.fr/libmsvm/data/data_exp/voc12/all/LIG_alex_goog_vgg_earlypw1.000p294pw0.610.svm|binary-labels]] and ii) [[http://mrim.imag.fr/libmsvm/data/data_exp/voc12/all/LIG_alex_goog_vgg_earlypw1.000p294pw0.610_mlabel.svm|multi-labels]]'	

  * Sub collections :Each subset contains a **select.txt** file, which indicates the keylist samples to be selected from the full collection.
      -  Developement set [[http://mrim.imag.fr/libmsvm/data/data_exp/voc12/dev/|Dev-set]]: divided into two subsets [[http://mrim.imag.fr/libmsvm/data/data_exp/voc12/train/|train-set]] and [[http://mrim.imag.fr/libmsvm/data/data_exp/voc12/val/|validation-set]]
      - Test set [[http://mrim.imag.fr/libmsvm/data/data_exp/voc12/test/|test]].