# Can gender categorization influence the perception of animated virtual humans?

**Authors:** V. Araujo, D. Schaffer, A. B. Costa, S. R. Musse  
**arXiv:** [2208.02386](https://arxiv.org/abs/2208.02386)  
**Date:** 2022-08-03  
**Summary page:** https://www.summarizepaper.com/en/paper/2208.02386v1/

## Key points

- Computer Graphics (CG) has enabled the creation of increasingly realistic animations, particularly in the representation of human models and behaviors through virtual humans.
- Gender is often stereotypically represented through movements, clothing, hair, and colors to ensure that virtual humans are understood by users as intended by designers.
- It is essential to understand whether participants' perceptions change depending on how a virtual human is visually presented.
- In a previous study, two groups of people watched a video of a real 9-month-old baby reacting to four stimuli while sitting in front of a mirror. The participants were separated into two groups: one group was told the baby was female, and the other group was told the same baby was male.
- To replicate this study using CG technology, researchers used a 3D model of a baby with animations for crawling, walking, and playing with a ball. Three stimulus objects were used for the virtual baby to interact with: a ball (positive reaction), jack-in-the-box (negative reaction), and colored unicorn (neutral reaction).
- A questionnaire was then distributed to participants who watched these videos.
- The results showed that people's emotional responses changed depending on the character gender attribute; in this case, only the difference in name affected their perception.
- The findings suggest that informing participants about the name of a virtual human can be enough to create gender perception that impacts their emotional response.
- This study highlights how CG technology can be used to investigate gender bias in relation to simulated babies and provides insight into how designers can guide modeling and animation of virtual humans to deliver the expected impact on the audience.

## Summary

The use of Computer Graphics (CG) has enabled the creation of increasingly realistic animations, particularly in the representation of human models and behaviors through virtual humans. However, gender is a characteristic that is often stereotypically represented through movements, clothing, hair, and colors to ensure that virtual humans are understood by users as intended by designers. Therefore, it is essential to understand whether participants' perceptions change depending on how a virtual human is visually presented. In a previous study, two groups of people watched a video of a real 9-month-old baby reacting to four stimuli while sitting in front of a mirror. The participants were separated into two groups: one group was told the baby was female, and the other group was told the same baby was male. The study aimed to assess whether the baby's gender influenced participants' perception of emotional and semantic levels. To replicate this study using CG technology, researchers used a 3D model of a baby with animations for crawling, walking, and playing with a ball. Three stimulus objects were used for the virtual baby to interact with: a ball (positive reaction), jack-in-the-box (negative reaction), and colored unicorn (neutral reaction). Three videos were created for each object as stimuli. A questionnaire was then distributed to participants who watched these videos. Statistical analyses were conducted using Kruskal-Wallis tests with Bonferroni correction as post hoc tests, Mann-Whitney tests, and Chi-Square tests through Python libraries Statsmodels and SciPy. The results showed that people's emotional responses changed depending on the character gender attribute; in this case, only the difference in name affected their perception. The findings suggest that informing participants about the name of a virtual human can be enough to create gender perception that impacts their emotional response. Overall, this study highlights how CG technology can be used to investigate gender bias in relation to simulated babies and provides insight into how designers can guide modeling and animation of virtual humans to deliver the expected impact on the audience.