XR2LEARN Emotion Recognition

A presentation of the Emotion Recognition concept, presented by Seyed Muhammad Hossein during MobileCHI Conference in Athens, September 2023.
XR2Learn's team explored the integration of emotion recognition technologies into virtual reality, highlighting the challenges in data collection, annotation, and multimodal integration. Also, we emphasized the potential of advanced learning techniques, like self-supervised and unsupervised learning, for better emotion interpretation. The paper advocates for innovative, personalized approaches in emotion data collection within VR environments. It underscores the unique opportunities VR presents for emotion recognition and calls for further research to harness its full potential. XR2Learn's team ultimately suggests a need for novel methodologies to address the complexities and fully exploit emotion recognition in VR settings.

Introduction:
Highlights the potential of integrating ER in VR.
Discusses unique challenges and opportunities in this field.
Background on Emotion Recognition:
Covers various ER modalities: physiological signals, facial expressions, voice, body language, and eye tracking.
Reviews existing ER datasets, differentiating between VR-specific and general datasets.
Challenges in ER for VR:
Addresses complexities in data collection, including eliciting and annotating emotional responses.
Discusses challenges in multimodal data integration and representation learning.
Opportunities in VR for ER:
Suggests innovative, personalized data collection methods.
Recommends self-supervised and unsupervised learning for data interpretation.
Conclusion:
Stresses the need for new methodologies in ER for VR.
Encourages further research to fully exploit ER's potential in VR.