One of the prominent clinical manifestations of schizophrenia is flat or altered facial activity, and flattening of emotional expressiveness (Flat Affect). In this study we used a structured-light depth camera and dedicated software to automatically measure the facial activity of schizophrenia patients and healthy individuals during a short structured interview. Based on K-means clustering analysis, facial activity was characterized in terms of Typicality, Richness and Distribution of 7 facial-clusters. Thus we found patients' facial activity to be poorer, more typical, and characterized mainly by neutral (flat) expressions. The facial features defined in our study achieved up to 85% correct diagnosis classification rate in a SVM based two-step algorithm, and were in significant correlation with Flat Affect severity. Our results demonstrate how the use of assistive technology and data-driven computational tools allow for a comprehensive description of patients' facial behavior in clinical settings, and may contribute to the reliability and accuracy of psychiatric diagnosis.
|Original language||American English|
|Title of host publication||3rd IEEE EMBS International Conference on Biomedical and Health Informatics, BHI 2016|
|Publisher||Institute of Electrical and Electronics Engineers Inc.|
|Number of pages||4|
|State||Published - 18 Apr 2016|
|Event||3rd IEEE EMBS International Conference on Biomedical and Health Informatics, BHI 2016 - Las Vegas, United States|
Duration: 24 Feb 2016 → 27 Feb 2016
|Name||3rd IEEE EMBS International Conference on Biomedical and Health Informatics, BHI 2016|
|Conference||3rd IEEE EMBS International Conference on Biomedical and Health Informatics, BHI 2016|
|Period||24/02/16 → 27/02/16|
Bibliographical notePublisher Copyright:
© 2016 IEEE.