J. Šalkevičius “Virtual reality for anxiety recognition using physiological signals – framework, methods and evaluation” doctoral dissertation defence

Author, Institution: Justas Šalkevičius, Kaunas University of Technology

Science area, field of science: Technological Sciences, Informatics Engineering, T007

Scientific Supervisor: prof. dr. Robertas Damaševičius (Kaunas University of Technology, Technological Sciences, Informatics Engineering, T007)

Dissertation Defence Board of Informatics Engineering Science Field:
Prof. Dr. Rimvydas Simutis (Kaunas University of Technology, Technological Sciences, Informatics Engineering, T007) – chairman,
Prof. Dr. Gennaro Cordasco (Luigi Vanvitelli University, Italy, Technological Sciences, Informatics Engineering, T007),
Prof. Dr. Arnas Kačeniauskas (Vilnius Gediminas Technical University, Technological Sciences, Informatics Engineering, T007),
Prof. Dr. Olga Kurasova (Vilnius University, Technological Sciences, Informatics Engineering, T007),
Assoc. Prof. Dr.  Simona Ramanauskaitė (Vilnius Gediminas Technical University, Technological Sciences, Informatics Engineering, T007).

The doctoral dissertation is available at the libraries of Kaunas University of Technology (K. Donelaičio St. 20, Kaunas, Lithuania) and Vilnius Gediminas Technical University (Saulėtekio al. 14, Vilnius)

Annotation:

This dissertation analyzes the virtual reality exposure therapy (VRET) systems for the treatment of anxiety disorders. VRET systems can help treat these disorders by placing the patients in a simulated 3D environment where they can gradually be exposed to the feared stimuli and learn to control their reactions. The anxiety-aware Virtual-Reality-as-a-Service (VRaaS) model is described in work. The proposed VRaaS model enables an adaptable VRET system with an integrated anxiety recognition component. This model allows the therapist to adapt virtual environments, change reactions of virtual avatars, and trigger various situations during treatment sessions to personalize the treatment session for a specific patient. The described VRET system features an integrated four-class anxiety recognition model. This model provides the therapist with the ability to recognize the patient’s anxiety during VRET at more distinct levels. Based on this capability to recognize anxiety, the therapist has more information on how to adapt the current virtual environment scenario for the patient in real-time (to lower or increase the intensity of the therapy). This thesis shows the fundamental knowledge of VRET system design, feature extraction, and classification of patients’ anxiety during VRET sessions when using machine learning methods. Finally, it presents an experiment with 30 participants conducted with a VRET system based on the proposed VRaaS model.