Representing human emotions has been a basic topic of research. The most usual emotion representation (ER) is the categorical one (7 basic categories). Discrete ER can also be described in terms of the FACS model, in which all possible facial actions are described in terms of Action Units (AUs). Finally, the dimensional model of affect -with valence and arousal (VA) being the most usual representations- has been proposed to distinguish between subtly different displays of affect and encode small changes in the intensity of each emotion on a continuous scale. This Competition is split into 3 Challenges: i) VA estimation, ii) 7 basic emotion classification, iii) AU detection. These Challenges use the Aff-Wild2, the first comprehensive benchmark for all 3 affect recognition tasks in-the-wild


Competition website: 


Competition start-ending dates:

  • 01 Nov 2019: Call for participation announced; start of Competition
  • 14 Feb 2020: Results and Paper submission deadline
  • 21 Feb 2020: End of review; Decisions sent to authors; Winner announcement
  • 28 Feb 2020: Camera ready versions deadline
  • 5 Mar 2020: End of Competition