Reproduce “FacePsy: An Open-Source Affective Mobile Sensing System -- Analyzing Facial Behavior and Head Gesture for Depression Detection in Naturalistic Settings” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Rahul Islam, Sang Won Bae
FacePsy, an open-source mobile system developed at Stevens Institute of Technology, enables privacy-preserving depression detection by analyzing facial behavior and head gestures directly on smartphones in naturalistic settings. The system achieves an 81% AUROC for classifying depressive episodes and an MAE of 2.81 for predicting PHQ-9 scores, identifying specific affective biomarkers.
Key insights
- Specific affective biomarkers, including certain Action Units (e.g., AU15, AU17), eye-open states, head pose (yaw), and smile expressions, are significant indicators for depressive episodes in naturalistic settings.
- Increased smiling probability showed a paradoxical positive correlation with depressive episodes, suggesting complex affective displays or phenomena like 'smiling depression'.
- Reliable depression detection is achievable with approximately one week of continuous data collection, with model performance reaching an AUROC of 0.714% by day seven.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.