Action Units Recognition by Pairwise Deep Architecture
This work addresses label inconsistency in facial expression analysis for competition settings, representing an incremental improvement.
The paper tackled the problem of Action Units (AUs) label inconsistency among subjects in the Affective Behavior Analysis in-the-wild (ABAW) competition by proposing a pairwise deep architecture, achieving a score of 0.67 compared to a baseline of 0.31 on the validation dataset.
In this paper, we propose a new automatic Action Units (AUs) recognition method used in a competition, Affective Behavior Analysis in-the-wild (ABAW). Our method tackles a problem of AUs label inconsistency among subjects by using pairwise deep architecture. While the baseline score is 0.31, our method achieved 0.67 in validation dataset of the competition.