The Action Similarity Labeling (ASLAN) Challenge

Selected Results:
For further information please see our paper.

  Method   Accuracy ± SE   AUC
  Human Performance 1   94.19   97.8
  ∑(x1'x2), HOG 1   56.58 ± 0.80   61.6
  ∑√(x1'x2), HOF 1   56.82 ± 0.57   58.5
  ∑√(x1'x2), HNF 1   58.87 ± 0.89   62.1
  12-sim, HOG+HOF+HNF 1   60.88 ± 0.77   65.3
  OSSML 2   64.25 ± 0.70   69.1
  MIP 3   64.62 ± 0.80   70.4
  MIP+HOG+HOF+HNF 3   65.45 ± 0.80   71.9

ROC Curves:

ROC curve

Creating ROC Curves:

To get access to the data, please fill in the ftp form in the downloads tab.

The following gnuplot script can be used to generate the ROC curves:
Create_ASLAN_ROC.p - link to ftp site

The script's input is a text file for each method.
Additional methods can be added to the script by adding their text files to the script command.
The lines in each text file are points on the ROC curve.
The format is:
[average true positive rate] [average false positive rate]

The current ROC files are available here: - link to ftp site


  1. O. Kliper-Gross, T. Hassner and L. Wolf
    The Action Similarity Labeling Challenge.
    IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 34(3), March 2012.

  2. O. Kliper-Gross, T. Hassner and L. Wolf
    The One Shot Similarity Metric Learning for Action Recognition.
    1st International Workshop on Similarity-Based Pattern Analysis and Recognition (SIMBAD), Sep. 2011

  3. O. Kliper-Gross, Y. Gurovich, T. Hassner and L. Wolf
    Motion Interchange Patterns for Action Recognition in Unconstrained Videos.
    12th European Conference on Computer Vision (ECCV), Oct. 2012
    [PDF] [Project and Code]

This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright. In most cases, these works may not be reposted without the explicit permission of the copyright holder.