Every "Body" Gets a Say: An Augmented Optimization Metric to Preserve Body Pose During Avatar Adaptation in...

Area of Science:

• Computer Science
• Human-Computer Interaction
• Virtual Reality

Background:

• User-avatar interaction in augmented reality (AR) is increasing, requiring avatar posture adjustments for tasks like object interaction or gaze maintenance.
• Current avatar redirection techniques often fail to preserve body posture, a crucial aspect of nonverbal communication.

Purpose of the Study:

• To develop a novel pose-preserving objective function for optimizing avatar kinematic configurations in AR.
• To enhance the realism and naturalness of avatar movements during user interactions.

Main Methods:

• Proposed a new objective function for multi-objective optimization of avatar kinematics.
• The function ensures correct joint placement and preserves joint orientation in space.
• Evaluated the approach against existing state-of-the-art redirection techniques.

Main Results:

• Achieved significant reduction in targeted redirection error compared to current methods.
• Demonstrated a simultaneous reduction in body posture error.
• Human subject testing confirmed more natural-looking redirection and realistic, believable avatar posture.

Conclusions:

• The new pose-preserving objective function effectively enhances avatar redirection in AR.
• This technique improves nonverbal communication fidelity in virtual interactions.
• Offers a more realistic and natural user experience in augmented reality applications.