The Impact of Avatar Retargeting on Pointing and Conversational Communication

Area of Science:

• Human-Computer Interaction
• Virtual Reality
• Computer Graphics

Background:

• Interacting with avatars in virtual reality (VR) allows users to embody different characters.
• Scaling avatars relative to users necessitates motion retargeting, impacting functional and social communication.
• Existing retargeting methods often preserve user pose or limb positions.

Purpose of the Study:

• To investigate how different motion retargeting strategies affect avatar pointing accuracy and spatial perception.
• To analyze the influence of retargeting on the social communication and perceived personality of third-person avatars.
• To explore the impact of avatar scaling and retargeting on user experience in VR.

Main Methods:

• Three studies were conducted focusing on third-person observed avatars.
• Established a baseline for near-field avatar pointing accuracy (approx. 5cm).
• Evaluated motion retargeting using direct pose transfer versus positional hand/ankle constraints, and Semantic Inverse Kinematics.

Main Results:

• Pointing accuracy is maintained with positional hand constraints but decreases with direct pose transfer.
• Semantic Inverse Kinematics can preserve accuracy by adjusting finger pointing direction.
• Retargeting methods slightly alter perceived avatar personality, impacting social communication, especially with significant proportion changes.

Conclusions:

• Motion retargeting significantly influences both functional (e.g., pointing accuracy) and social (e.g., perceived personality) aspects of avatar communication in VR.
• Positional constraints offer a balance between functional accuracy and social perception.
• The impact of retargeting on social communication may be task-dependent and influenced by avatar proportion differences.