Review of VREED in Eye Tracking-Based VR Emotion Detection

Eye-tracking emotion detection, especially within virtual reality (VR) environments, has potential to be a key technology that has implications across various industries. The VREED dataset serves as a baseline for research in emotion classification tasks using eye-tracking, galvanic skin response (GSR), and electrocardiogram (ECG) data. This study expands on the original VREED research by applying 17 different classifiers across six emotion classification tasks. Feature selection is used to address data scarcity and assess the impact of selecting specific features versus utilizing the full feature set. Results showed some promising classifiers, with feature selection further improving performance for certain models. This demonstrates the value of optimizing this feature selection process and suggests further techniques in yielding better results. This study serves as an extension of the baseline of VREED and provides insight into improving the robustness of emotion detection models in VR, with applications in healthcare, education, and gaming.