Fully Expressive Avatar in VR Collaboration

Spring 2023, Dave Song

Using Fully Expressive Avatar to Collaborate in Virtual Reality: Evaluation of Task performance, presence, and attraction 

Motivation: Interested in developing standard evaluation metrics for VR collaboration, I was also interested in exploring avatar as a major factor in the quality of UX. Wanted to explore if the highly expressive avatar developed in the paper had any significant difference.

<Abstract>

• Avatar-mediated collaboration in a virtual environment

• Study of highly expressive avatar:

  • definition of a highly expressive avatar they provide is one that can provide high levels of nonverbal expression by tracking behavior such as body movement, hand gestures, and facial expression.

• With the avatar-mediated vr system, participants were asked to perform tasks.

• Highly expressive avatar

  • more social presence and attraction

  • better task performance

• compared to low-expressive counterparts. → vr system can benefit from the high level of nonverbal expressiveness.

<Introduction>

• avatar realism: the measure of avatar quality

• appearance

• behavioral realism

• People tend to communicate more through nonverbal behavior during social interaction compared to verbal channels.

  • • Matumoto et al., 2012

• performed Charades game in the shared virtual environment(they call it SVE).

  • • rational behind choosing charades: engaging with nonverbal behavior to complete a collaborative challenge

• The main contribution of the research

  • fully expressive avatar control system

    • supports eye-gaze

    • mouth rendering combined with tracking natural nonverbal behavior

      • using multiple LEAP motion tracking cameras

  • evaluation of the effects of different levels of expression through an avatar on communication and collaboration.

<Key Points from Related Work Section>

• quality of SVE depends on whether it can support social interaction. This requires users’ appearances and behavior and motion

  • nonverbal cues

  • quality of embodiment → higher social presence. “can lead to higher social presence ratings compared to face-to-face interactions

• Previous research is limited in that they omitted or did not fully support the nonverbal behaviors when examining collaboration between users

<Avatar control system and representation>

• conventionally 3 tracking points with HMD, and two controllers. → mostly just floating avatars

• Use of RGB-D sensor and VR device as a potential solution

• Additionally, hand gestures could be tracked by using Leap Motion Controller.

• Eye gaze and facial expression

  • RGB-D sensor to track facial expressions and eye gaze

<Technical Setup>

• Body Tracking could be done by using 4 Kinect v2 devices located at the four corners of the tracking area

• Hand tracking was done using a multi-LMC system where they had five LMCs installed on the HMD

• Eye gaze direction: generated random gaze shift to mimic natural eye movement.

<Measurements>

objective and subjective data were collected. Participants completed questionnaires every two sessions after they performed the charades as a performer and a guesser.

Results of HEA vs. LEA

• Co-presence: “feeling that the user is with other entities”

  • did not show any significant difference between HEA and LEA

• Social Presence: “feeling of the user, which makes people feel connected with other through the telecommunication system”

  • a significant difference between the two

• Interpersonal Attraction: “the measurement of liking and attraction”

  • also, HEA made a significant difference

Source: https://www.frontiersin.org/articles/10.3389/frvir.2021.641296/full