Attention-VR: predicting VIMS from attention-weighted multi-modal movement data

E. Chang , E. Lim , H. Cho , J. D. Hart , A. Dey , Z. Zhang , H. Yang , M. Billinghurst

ISMAR-Adjunct

Attention-VR predicts visually induced motion sickness in VR by integrating head, eye, and mouth movement data and dynamically giving more weight to whichever of these signals is most informative for spotting early sickness. The model was built from data collected on 38 participants in a custom VR rollercoaster. Combining the movement signals this way improves both prediction accuracy and interpretability, clarifying how much each modality contributes to the earliest signs of discomfort.

@inproceedings{chang2025,
  author    = {E. Chang and E. Lim and H. Cho and J. D. Hart and A. Dey and Z. Zhang and H. Yang and M. Billinghurst},
  title     = {Attention-VR: predicting VIMS from attention-weighted multi-modal movement data},
  booktitle = {ISMAR-Adjunct},
  pages     = {664-665},
  doi       = {10.1109/ISMAR-Adjunct68609.2025.00138},
  year      = {2025}
}