Emotion-specific AUs for micro-expression recognition

Abstract The Facial Action Coding System (FACS) comprehensively describes facial expressions with action units (AUs). It is a well-used technique by researchers in emotions research to understand human better. Most micro-expression datasets provide FACS-coded AU ground truths corresponding micro-expressions classes. commonly accepted computer vision-based that certain are reliably revealed when specific combinations of AUs occur. However, the reliability truth lower than normal expressions, as they have intensities. Moreover, these only report overall all AUs. could not be identified which had been accurately coded. This work aims revisit popular datasets, namely CASME II, SAMM and CAS(ME) $$^2$$ 2 , inspect whether any crucial for recognition may need reconsidered. paper also provides detailed analysis yields new AU-based RoIs each dataset. These improve performances compared baselines considered this work. proposed rates $$2\%$$ % $$1\%$$ 1 $$4\%$$ 4 respectively, existing RoIs.