Modeling Dynamics of Facial Behavior for Mental Health Assessment
This work addresses mental health assessment by providing a novel method for analyzing facial behavior, though it is incremental as it adapts existing techniques to a new domain.
The study tackled the sparse representation of facial action units by modeling facial expression dynamics using NLP-inspired embeddings, achieving improved performance in schizophrenia symptom estimation and depression severity regression over baseline FAU-only models.
Facial action unit (FAU) intensities are popular descriptors for the analysis of facial behavior. However, FAUs are sparsely represented when only a few are activated at a time. In this study, we explore the possibility of representing the dynamics of facial expressions by adopting algorithms used for word representation in natural language processing. Specifically, we perform clustering on a large dataset of temporal facial expressions with 5.3M frames before applying the Global Vector representation (GloVe) algorithm to learn the embeddings of the facial clusters. We evaluate the usefulness of our learned representations on two downstream tasks: schizophrenia symptom estimation and depression severity regression. These experimental results show the potential effectiveness of our approach for improving the assessment of mental health symptoms over baseline models that use FAU intensities alone.