Readspeaker menu

Jena Action Recognition Dataset


The Jena Action Recognition Dataset is a collection of videos showing remotely triggered SONY ERS-7 Aibo robot dogs for evaluating and benchmarking model-free action recognition systems.

Here you are able to download the complete dataset including:

  • videos of all sequences
  • background images
  • calibration data
  • bounding boxes


If you use this data, we kindly ask you to cite the following paper:

Marco Körner and Joachim Denzler. JAR-Aibo: A Multi-View Dataset for Evaluation of Model-Free Action Recognition Systems. Proceedings of the 1st International Workshop on Social Behaviour Analysis (SBA). 2013. p. 527-535

If you compare to our baseline results, we even more kindly ask you also cite also the following paper:

Marco Körner and Joachim Denzler. Temporal Self-Similarity for Appearance-Based Action Recognition in Multi-View Setups. Proceedings of the 15th Conference on Computer Analysis of Images and Patterns (CAIP). 2013. p. 163-171

Locations of visitors to this page