Mixing in Reverse Optical Flow to Mitigate Vection and Simulation Sickness in Virtual Reality

Document Title

Mixing in Reverse Optical Flow to Mitigate Vection and Simulation Sickness in Virtual Reality

Document Information

• Subject Area: Virtual Reality (VR), Motion Sickness Mitigation, Human-Computer Interaction
• Keywords: Virtual Reality, Sensory Conflict, Simulation Sickness, Optical Flow, Reverse Optical Flow Visualization, Immersion, Sensory Disruption, Vection

Research Background and Problem

• Problem or Challenge:

  • Simulation sickness (simulator sickness or VR sickness) is one of the main barriers to the widespread adoption of virtual reality applications, particularly in virtual navigation scenarios.
  • The primary cause of simulation sickness is the conflict between visual and vestibular sensory inputs (sensory mismatch), where users receive visual motion cues from the environment (i.e., vection) while remaining physically stationary.
  • Traditional solutions (e.g., reference objects, blurring, field-of-view adjustments) are partially effective but often compromise content fidelity and the immersive experience.

• Significance of the Research:

  • Addressing simulation sickness can enhance the acceptance of VR technology and improve user experience quality, reducing potential negative physiological and psychological impacts.

• Motivation and Related Work:

  • Visual optical flow is a key source of visual motion perception. Studies [7][14][15] have shown that manipulating optical flow characteristics can regulate vection and related motion sickness.
  • The concept of reverse optical flow has not been thoroughly explored in the context of mitigating VR sickness.
  • The authors hypothesize that overlaying reverse optical flow may reduce sensory conflicts and alleviate simulation sickness.

Solution

• Proposed Method:

  • Compute and visualize reverse optical flow of virtual visual motion to reduce sensory conflicts and alleviate simulation sickness.

• Innovations:

  • Introduced a novel visual mixing method based on reverse optical flow, aiming to directly reduce sensory conflicts caused by virtual visual motion.
  • Compared the mitigating effects of reverse optical flow under varying degrees of freedom (DOFs) of motion and visual complexity.

• Implementation Steps and Key Techniques:

  1. Reverse Optical Flow Algorithm:
     • Used the Lucas-Kanade optical flow algorithm and Shi-Tomasi feature detection algorithm to compute reverse optical flow.
     • Implemented dynamic visualization by overlaying reverse optical flow as short line segments in the scene.
  2. Experimental Design:
     • Designed primary and auxiliary experiments using unified content or other typical navigation content.
     • Quantified simulation sickness using the Simulator Sickness Questionnaire (SSQ) and evaluated user immersion and experience with an improved SUS questionnaire.
  3. Algorithm Performance:
     • The optical flow algorithm runs in real-time, achieving an interactive rate of 80-90 fps, suitable for VR environments.

Research Findings

• Specific Results:

  • Reverse optical flow significantly reduced simulation sickness, particularly under 2DOF, 3DOF, and 5DOF motion conditions.
  • In specific content (e.g., "Ship Ride" scenario), the effectiveness of the method may be diminished by optical flow characteristics, such as low visual complexity.
  • Low-density optical flow may lead to insufficient suppression effects, while overly high density may negatively impact user experience.

• Comparison with Existing Solutions:

  • Compared to traditional methods, reverse optical flow directly addresses the root cause of sensory conflict and demonstrated significant advantages in multiple DOF tests.
  • However, compared to certain methods (e.g., dynamic field-of-view adjustment), reverse optical flow may introduce higher content interference.

• Experimental or Evaluation Results:

  • Tested the sickness mitigation effects under different DOF conditions and typical navigation content.
  • Significant mitigation effects under different DOF conditions were as follows: 2DOF (p = 0.013), 3DOF (p = 0.003), 5DOF (p = 0.003), with 4DOF showing no significant results.
  • In auxiliary experiments, content such as "Rollercoaster" and "Car Race" demonstrated significant effects, while "Ship Ride" showed no significant improvement.

• Limitations and Future Directions:

  • Limitations:
    • Many algorithm parameters were empirically determined, lacking generalizability.
    • The current algorithm only partially alleviates sensory conflicts.
    • Reverse optical flow showed limited performance in low visual complexity scenarios (e.g., ocean backgrounds).
  • Future Directions:
    • Investigate the relationship between reverse optical flow and kinematic characteristics (e.g., linear/rotational motion) as well as visual complexity.
    • Explore localized distributed optical flow mixing strategies and dynamic triggering methods for reverse optical flow.
    • Combine with existing motion sickness prediction technologies to dynamically adjust reverse optical flow in response to sudden motion changes.
    • Explore alternative forms of reverse optical flow visualization (e.g., color coding, shape markers, edge information).

Conclusion

This paper pioneers a reverse optical flow mixing method to mitigate simulation sickness, validates its effectiveness, and identifies several future research directions. Despite certain limitations, this method provides a novel approach to addressing motion sickness in the field of virtual reality technology.