-
1
Automated emotion recognition systems often rely on acted datasets, which overlook the nuances of spontaneous emotional expressions.
-
2
A large corpus of authentic facial emotion expressions was created from outpatient psychotherapy sessions, annotated with free-text descriptions.
-
3
Model predictions achieved a mean z-score of 1.97 in matching human annotations, indicating strong performance in emotion recognition.
-
4
AFFECT, an open-source pipeline, was developed to analyze emotional expressions in everyday video recordings using advanced machine learning techniques.
-
5
The study highlights the complexity of human emotion recognition, which is influenced by context, culture, and personal characteristics.
