Automatic Phobia Detection with Virtual Reality and Machine Learning Algorithms

Anxiety is a prevalent mental health condition characterized by persistent feelings of fear, worry, and unease. It can manifest in various forms, such as generalized anxiety disorder, social anxiety, panic disorder, or specific phobias. Individuals with anxiety disorders often experience physiological symptoms, such as increased heart rate, sweating, trembling, and difficulty concentrating. Virtual Reality (VR) creates a computer-generated, interactive, and immersive environment that simulates real-world scenarios with a high degree of realism. The immersive nature of VR allows individuals to experience realistic scenarios and elicits emotional and physiological responses like real-life situations. The incorporation of evidence-based therapeutic techniques, coupled with accurate patient identification through machine learning, ensures a personalized and targeted treatment approach. In this paper, system-based VR and Artificial Neural Network (ANN) based machine learning algorithm are utilized for phobia detection. The proposed system consists of two main phases; front-end and back-end. In the front-end phase, several scenes are developed in VR to simulate three different levels of phobia. In this phase, the microphone hardware is utilized, where the sound of the player is recorded. This recorded voice signal is analyzed in the back-end phase. In this phase, ANN is utilized for phobia detection. The proposed system is quite promising, according to the experimental findings. It obtained an overall accuracy of 97% for the collected dataset.